# Everpod

> Everpod is an easy way to get your own always-on, persistent cloud computer for AI agents, working in minutes: with a managed OpenClaw agent on it, or as a developer pod with Claude Code and Codex installed. One customer per machine. An OpenClaw pod is a managed OpenClaw agent on a private computer of its own, set up, secured, backed up, and kept up to date, for $29/mo (the first month half price) with up to $10 of model usage included each month. A developer pod is a cloud computer you run yourself, with Claude Code and Codex installed, reached only over your own Tailscale network, for $24/mo with no model usage included.

Every pod has at least 2 vCPUs, 4 GB of RAM and 40 GB of storage.

You can create an OpenClaw agent at https://everpod.ai/create and a developer pod at https://everpod.ai/create/developer, or have an agent you already use start either for you through the Everpod API (below). The pages below cover hosting, running, and securing AI agents, what managed hosting includes, and what people actually do with an agent that has its own computer.

## Product

- [Everpod — your own open-source cloud agent, in minutes](https://everpod.ai/): Managed OpenClaw hosting. Your own open-source AI agent gets a pod: a private, always-on cloud computer of its own, set up, secured and backed up for you. $29 a month, model usage included.
- [What a cloud agent can do — real AI agent use cases](https://everpod.ai/cloud-agent-use-cases): Real things people do with AI agents on their own always-on computer: morning briefings, standing jobs, overnight research, watching, remembering.
- [OpenClaw hosting, explained — hosted pods vs DIY](https://everpod.ai/hosted-openclaw): Hosted OpenClaw: your agent runs on a managed, always-on cloud machine, not your laptop or a DIY server. What a host does, what stays yours, what to look for.
- [Open-source AI agent hosting — a plain-language explainer](https://everpod.ai/ai-agent-hosting): Open-source AI agents like OpenClaw need an always-on, secure computer to live on. What agent hosting means, the options, and what a good setup includes.
- [OpenClaw guides — setup, hosting, security](https://everpod.ai/guides): OpenClaw guides: setup, the chat apps, models, where your agent should live, updates and backups, security, and agents on the web, from the team that runs OpenClaw pods.
- [Developer pod: your own always-on cloud computer for coding agents](https://everpod.ai/developer-pod): An always-on cloud computer of your own with Claude Code and Codex installed, reached only over your own Tailscale network. 4 vCPU and 8 GB for $24 a month. You administer the machine and sign in to each agent with your own subscription or API key; no model usage is included. Start one at https://everpod.ai/create/developer, or have a coding agent on your own computer start it and help you connect it (https://everpod.ai/docs/api#developer-pod, under "Get a developer pod with your agent").

## Trust and legal

- [About Everpod](https://everpod.ai/about): Who is behind Everpod: David Harding, who builds and runs it through his Wyoming software company, DEH Technologies LLC, which is also behind Invoice Data Extraction.
- [Security overview](https://everpod.ai/security): How your pod and your account are protected: one machine per customer, no incoming connections unless you switch on a public address, daily backups, permanent deletion when you leave, and who can reach what.
- [Privacy Policy](https://everpod.ai/privacy-policy): What Everpod holds about you, where it lives, how long it's kept, and the plain-language answer to who can see what's on your pod.
- [Terms of Service](https://everpod.ai/terms-of-service): The binding agreement: the service, billing, acceptable use, your content and when we access your pod, warranties and liability.
- [Refund and Cancellation Policy](https://everpod.ai/refund-policy): How cancellation works, when your pod is deleted, and when a refund applies.
- [Everpod status](https://status.everpod.ai/): Whether Everpod is working right now, and day by day over the last 90 days (the history as data: https://status.everpod.ai/history.json).

## Signing up, signing in, and reaching us

- Creating an agent at https://everpod.ai/create is the signup: there is no separate account step and no password. Three steps: name your agent; confirm your email with the six-digit code we send you, then pay by card, Apple Pay or Google Pay (prices in US dollars, shown in your own currency where supported, tax added where it applies); about 15 minutes later your agent is awake and we walk you through connecting a messaging app.
- Returning owners sign in at https://everpod.ai/signin with their email and a code we send; on a device they have used before they land in their account without seeing the sign-in page.
- The signup at https://everpod.ai/create sells the $29 plan only. Bigger plans start at $59 a month and are switched from the pod's own page once it is running. When the included usage is used up, the agent's thinking on it pauses until the plan renews; a bigger plan or the owner's own model key keeps it going, and Everpod adds nothing on top of an owner's own key.
- Cancel anytime from your account, or by emailing support@everpod.ai from the account's email. The pod runs to the end of the paid period and is permanently deleted 7 days after that, backups included.
- Contact: support@everpod.ai for anything about the service, privacy@everpod.ai for privacy, security@everpod.ai for security reports. Everpod is operated by DEH Technologies LLC, a Wyoming (United States) company.

## The Everpod API

- [Everpod API](https://everpod.ai/docs/api): How to connect an agent you already use, such as Claude Code or Codex, to your Everpod account, and have it start an OpenClaw pod or a developer pod for you: making a key, the MCP server, the REST routes, and each pod status.
- A key lets an agent or an app you trust see your pods and start a new one for you, which you then pay for on everpod.ai. It can’t pay, change or cancel a plan, delete anything, open your agent’s control panel, or reach a developer pod’s machine.
- An agent starts either kind of pod. For a developer pod, a coding agent on your own computer, such as Claude Code or Codex, starts the machine and tells you each step that is yours: you make the key, pay, and connect the machine to your Tailscale network from its page. The steps are at https://everpod.ai/docs/api#developer-pod, under "Get a developer pod with your agent".
- For an agent that takes skills, the same steps are a skill, everpod-developer-pod, from https://github.com/everpod-ai/skills. Claude Code installs it with claude plugin marketplace add everpod-ai/skills and then claude plugin install everpod-developer-pod@everpod; Codex with codex plugin marketplace add everpod-ai/skills and then codex plugin add everpod-developer-pod@everpod. For another agent, copy its folder into the agent's skills folder.
- An agent connects to the MCP server at https://everpod.ai/mcp, and a script calls the REST routes under https://everpod.ai/api/v1, each sending the key as the header "Authorization: Bearer <key>". The account's owner makes and revokes keys at https://everpod.ai/account/keys, signing in there with their email address and the code we send; for an owner with no account yet, signing in makes one.
- An OpenClaw agent can install the Everpod skill from ClawHub (https://clawhub.ai/everpod/skills/everpod), and it connects itself, starts a new agent or a developer pod for its owner, and can move itself onto a pod: openclaw skills install @everpod/everpod
- Client libraries make the same three calls: @everpod-ai/sdk on npm for JavaScript (https://www.npmjs.com/package/@everpod-ai/sdk) and everpod-ai on PyPI for Python (https://pypi.org/project/everpod-ai/).

## Guides: Getting started

- [What are OpenAI dots? ChatGPT's always-on agents, explained](https://everpod.ai/guides/what-are-openai-dots): OpenAI's dots are always-on ChatGPT agents with their own cloud computer. Which plans include one, what it costs, what it can reach, and what stays with you.
- [What is Meta Muse? Meta's personal AI agent, explained](https://everpod.ai/guides/what-is-meta-muse): Meta's personal AI agent on a cloud computer of its own, launched September 8, 2026: what it can reach, when it asks, what it costs, and which Muse is which.
- [What is AGENTS.md, and does your agent actually read it?](https://everpod.ai/guides/what-is-agents-md): AGENTS.md is the open format for an agent's standing instructions. Which agents load it, how nested files combine, and whether files it points to get read.
- [Keep your AI agent's personality and memory in files you own](https://everpod.ai/guides/keep-your-agents-personality-and-memory): An OpenClaw agent's character and memory are text files on its computer, so a model switch or shutdown can't take them. What changes, what to keep, how to move.
- [Run OpenClaw with Ollama](https://everpod.ai/guides/openclaw-ollama): Connect OpenClaw to local or remote Ollama models. Choose the right endpoint, check tool support, and distinguish local inference from Ollama Cloud.
- [Where to get OpenClaw: the official site, docs, and downloads](https://everpod.ai/guides/official-openclaw-sources): openclaw.ai, docs.openclaw.ai, the GitHub org, and ClawHub are the real sources, and a crowd of lookalike sites isn't. How to verify what you're installing.
- [What are OpenClaw skills, and how do they work?](https://everpod.ai/guides/what-are-openclaw-skills): Skills are instruction packs that teach your agent repeatable workflows. How they load, where they come from, and how to add them without importing risk.
- [What is ClawHub? OpenClaw's skill registry, explained](https://everpod.ai/guides/what-is-clawhub): ClawHub is the public registry where OpenClaw skills and plugins are shared. How installing works, and the supply-chain caution every user should know.
- [Running OpenClaw on Windows: native vs WSL2](https://everpod.ai/guides/openclaw-on-windows): OpenClaw runs on Windows both natively (as a Scheduled Task) and under WSL2, and the project itself calls native Windows the trickier path. How to choose.
- [Can you run OpenClaw on a Raspberry Pi?](https://everpod.ai/guides/openclaw-raspberry-pi): Yes: official ARM images exist, and a Pi makes a charming always-on agent box. The caveats: RAM, storage wear, and skills that assume x86.
- [Hatching your OpenClaw agent: names, identity, and personality](https://everpod.ai/guides/openclaw-hatching): OpenClaw makes a small ritual of an agent's first boot. What hatching actually sets up, why it matters, and how to shape who your agent becomes.
- [What is OpenClaw? The open-source AI agent, explained](https://everpod.ai/guides/what-is-openclaw): OpenClaw is free, open-source software that turns an AI model into a personal agent you message from WhatsApp or Telegram. What it does, how it works, and what it takes to run.
- [How much does OpenClaw cost? (It's free — here's what isn't)](https://everpod.ai/guides/how-much-does-openclaw-cost): OpenClaw is free. Compare hosting, model and connected-tool costs with a dated VPS price, worked monthly budgets and what an existing ChatGPT account can cover.
- [How to install OpenClaw: the three commands, and what comes after](https://everpod.ai/guides/install-openclaw): Installing OpenClaw takes minutes with npm. What you need first, what each command does, and the setup decisions that actually matter.
- [Running OpenClaw in Docker: images, volumes, and the port-binding trap](https://everpod.ai/guides/openclaw-docker): The official Docker setup, what must be volume-mounted for your agent to survive, and the default that quietly puts the Gateway on your network.
- [The OpenClaw dashboard (Control UI): what it is and how to open it](https://everpod.ai/guides/openclaw-dashboard): OpenClaw ships a web dashboard for chatting with and managing your agent. Where it lives, how sign-in works, and how to reach it safely from anywhere.
- [How OpenClaw memory works — and how to see what your agent knows](https://everpod.ai/guides/openclaw-memory): Your agent's memory is plain files you can open, edit, and back up. What OpenClaw remembers, where it keeps it, and how to manage it.
- [What is the OpenClaw Gateway?](https://everpod.ai/guides/what-is-the-openclaw-gateway): The Gateway is the always-on process at the heart of OpenClaw: it connects your chat apps to your agent. Here's what it does, in plain language.
- [Do you need an API key to use OpenClaw?](https://everpod.ai/guides/do-you-need-an-api-key-to-use-openclaw): Yes: OpenClaw brings the harness, you bring the model. What a model key is, where to get one, and what it costs.
- [Where does OpenClaw store its config, memory, and workspace?](https://everpod.ai/guides/where-does-openclaw-store-its-data): OpenClaw keeps its config, sessions, and your agent's memory under one home directory. Where it lives, what's inside, and what to back up.
- [Who created OpenClaw? The story behind Clawdbot, Moltbot, and the Foundation](https://everpod.ai/guides/who-created-openclaw): Peter Steinberger built OpenClaw as a side project in late 2025. The viral rise, the two renames, the OpenAI move, and who runs the project now.
- [Is OpenClaw legit and safe to install?](https://everpod.ai/guides/is-openclaw-legit): Yes. MIT-licensed, foundation-governed, developed in the open. Where the real risks live instead: lookalike sites, third-party skills, and where you run it.
- [Using your ChatGPT subscription with OpenClaw (Codex sign-in)](https://everpod.ai/guides/openclaw-chatgpt-subscription): Connect ChatGPT to OpenClaw, choose a model your account supports, and check what happens at usage limits. Subscription access and API billing are separate.
- [Can you use your Claude subscription with OpenClaw?](https://everpod.ai/guides/openclaw-claude-subscription): Anthropic banned it in April 2026, then walked it back twice. What its terms say now, why agent workloads exhaust subscription limits, and the predictable alternatives.
- [Which ChatGPT plans include Codex? All of them — here's the breakdown](https://everpod.ai/guides/which-chatgpt-plans-include-codex): Every ChatGPT plan includes Codex, Free included. Plus limits for Astra, GPT-6.1 Sol and Luna, Pro at $100, $200 or $500, and Business Premium seats.
- [The ChatGPT setting that makes device-code sign-in work](https://everpod.ai/guides/codex-device-code-authorization): Enable Codex device-code login in ChatGPT Security settings or workspace permissions, start a fresh sign-in, and distinguish account, code and model errors.
- [The best model for OpenClaw: what we run, and the default trap](https://everpod.ai/guides/best-model-for-openclaw): Fresh OpenClaw setups pick GPT-6 Astra: $10/M input, and scarce on a Plus plan. Why agent bills are input bills, with GPT-6.1 Sol, Luna and their prices.
- [SOUL.md, IDENTITY.md, USER.md: the files that make your agent](https://everpod.ai/guides/openclaw-workspace-files): OpenClaw's identity and memory are plain markdown injected into the system prompt on every run. What each file does, the truncation budgets, and why editing them is the point.
- [OpenClaw plugins vs skills: code, instructions, and the trust line](https://everpod.ai/guides/openclaw-plugins): Skills are markdown; plugins are code running inside your Gateway, and most channels are plugins. The allowlist mechanics, and what that startup warning means.
- [AI agent memory, explained: what your agent remembers and where it lives](https://everpod.ai/guides/ai-agent-memory): An agent's memory is files, indexes, and transcripts on a disk somewhere, not something inside the model. The three layers, the versioned-git pattern, and why the disk's location matters.
- [What is Hermes? Nous Research's self-improving open-source agent](https://everpod.ai/guides/what-is-hermes-agent): Hermes is an MIT-licensed agent that lives on your server, writes its own skills, and works with any model, local ones included. What ships today, how it installs, and the OpenClaw relationship.
- [“Please contact your workspace admin to enable device code authentication” — the fix](https://everpod.ai/guides/chatgpt-workspace-device-code): A Codex sign-in error asking for your workspace admin is a permission issue. What the admin enables, how to retry, and supported alternatives if it stays off.
- [Which Node.js version does OpenClaw need?](https://everpod.ai/guides/openclaw-node-version): Node 22.22.3+, 24.15+, or 25.9+, with 26 the recommended default as of mid-2026. How to check yours, the npm --allow-scripts wrinkle, and updating Node without breaking other projects.
- [Build OpenClaw from source, or run the prebuilt image?](https://everpod.ai/guides/openclaw-source-vs-docker): npm global, prebuilt container, or git clone + pnpm: three official install paths with different trust and effort profiles. Who each one is for, and what the Docker setup script does by default.
- [What is a Docker container? Images, layers, and where data goes](https://everpod.ai/guides/what-is-a-docker-container): A container is an isolated process built from a read-only image plus one writable layer, and it forgets everything when deleted. Why that design is right for software that must not forget.
- [What does docker compose up -d actually do?](https://everpod.ai/guides/docker-compose-up-d): It builds, creates, and starts everything your compose.yaml describes, detached into the background. What -d changes, stop vs down, and who restarts the container when it dies.
- [What does npm install -g actually do?](https://everpod.ai/guides/npm-install-g): The -g flag installs a package once for the whole machine and puts its command on your PATH. Where globals live, the EACCES fix, and npm 12's install-script approval flag.
- [How to install WSL2 (and what it actually is)](https://everpod.ai/guides/install-wsl2): One admin PowerShell command installs a real Linux kernel in a lightweight VM, Ubuntu included. The install steps, the BIOS trap, the systemd detail, and what it means for agents.
- [What openclaw onboard actually does, step by step](https://everpod.ai/guides/openclaw-onboard): The setup wizard proves a model works with a real completion before it builds anything else. QuickStart vs Manual, the daemon choice, where files land, and where hatching begins.
- [What is Grok Bot? xAI's always-on AI teammates, explained](https://everpod.ai/guides/what-is-grok-bot): Named, persistent bots that sign into your tools and keep working from one shared cloud computer per account. What shipped in the August 2026 beta, what it costs, and the boundary to understand.
- [Which AI models can a hosted OpenClaw agent use, and how do you switch?](https://everpod.ai/guides/hosted-openclaw-models): Essentially all of them: the latest from DeepSeek, Qwen, OpenAI, Anthropic, Google, and more, switched with a sentence in chat. The two levers that work, and why a switch might seem not to land.
- [How to start a new conversation with OpenClaw — and what it remembers](https://everpod.ai/guides/openclaw-new-session): /new archives and starts fresh, /reset resets in place, and workspace memory survives both. Plus 2.0's quiet change: the daily 4 a.m. auto-reset is gone, so sessions now run until you end them.
- [What is Bot Mode for Hermes? Named bots, explained](https://everpod.ai/guides/hermes-bot-mode): Hermes Desktop turns agent profiles into a roster of named Bots, each with its own role, model, memory, and skills, messaging each other by @mention. How it works and what it costs to run.
- [What is OpenClaw dreaming? DREAMS.md and the 3 a.m. memory run, explained](https://everpod.ai/guides/openclaw-dreaming): OpenClaw 2 consolidates memory overnight: light, REM and deep phases at 03:00 by default, a Dream Diary in DREAMS.md, promotions into MEMORY.md. What it does, what it costs, and the keys that tune or stop it.
- [How to change OpenClaw's thinking level (reasoning effort), for one chat or for good](https://everpod.ai/guides/openclaw-thinking-level): /think sets the reasoning effort for the conversation you are in; agents.defaults.thinkingDefault sets the install-wide default, and unset it lands on medium. The levels, the resolution order, and how to check what actually ran.
- [Why your OpenClaw agent can't install anything in Docker, and what survives an image update](https://everpod.ai/guides/openclaw-docker-missing-tools): The official image runs as a non-root user with a lean toolset, so apt is out and anything installed into the container layer dies at the next pull. OPENCLAW_IMAGE_APT_PACKAGES, a persistent home volume, and what the docs say survives.
- [Do you need to know how to code to use OpenClaw?](https://everpod.ai/guides/do-you-need-to-code-to-use-openclaw): No code, but some computer work: installing it yourself means a terminal, a model sign-in and a chat-app token, and running it always-on means a server. Where the technical work actually sits, where non-developers stall, and the route with nothing to install.
- [What is Squad? AI teammates on a dedicated OpenClaw computer, explained](https://everpod.ai/guides/what-is-squad): Squad launched September 3, 2026: named AI teammates on one dedicated cloud computer per customer, on the ChatGPT or Claude plan you already pay for, $99 a month. What it runs on, what it does not publish, and how it differs from Grok Bot.
- [What is Hermes Desktop? The app, the agent and remote connections](https://everpod.ai/guides/what-is-hermes-desktop): Hermes Desktop puts a graphical interface around Hermes Agent. What it adds, how remote connections differ from hosting, and why remote chat is not computer control.

## Guides: Channels

- [Connecting OpenClaw to Telegram, step by step](https://everpod.ai/guides/openclaw-telegram): Telegram is the easiest first channel for OpenClaw: a bot token from BotFather, one config block, no exposed ports. The full setup, including pairing.
- [Connecting OpenClaw to WhatsApp: how it works and what to expect](https://everpod.ai/guides/openclaw-whatsapp): OpenClaw joins WhatsApp the way WhatsApp Web does: a QR link, no Business API. Setup steps, the second-number question, and the gotchas.
- [Connecting OpenClaw to Discord, step by step](https://everpod.ai/guides/openclaw-discord): Create the bot, enable the Message Content intent, invite it to your server, and hand OpenClaw the token. Plus the allowlist trap and how memory behaves in servers.
- [Connecting OpenClaw to Slack (Socket Mode, two tokens, no exposed ports)](https://everpod.ai/guides/openclaw-slack): The app manifest, the bot token vs app-level token confusion, pairing for DMs, and the channel-ID gotcha that silently blocks messages.
- [Connecting OpenClaw to Signal with signal-cli](https://everpod.ai/guides/openclaw-signal): Signal has no bot API, so your agent gets its own number. Dedicated number vs linked device, the registration steps, and the deauth trap to avoid.
- [OpenClaw and iMessage: what it really takes](https://everpod.ai/guides/openclaw-imessage): iMessage needs a signed-in Mac running the imsg CLI, and the fancy features need SIP disabled. What works at each stage, and whether it's worth it.
- [OpenClaw voice calls: giving your agent a phone number](https://everpod.ai/guides/openclaw-voice): The @openclaw/voice-call plugin places and answers real phone calls via Twilio, Telnyx, or Plivo. Setup, the realtime-vs-streaming choice, how it differs from voice notes in chat, and the install form that matches a pinned OpenClaw.
- [Connecting OpenClaw to Google Chat: webhooks, service accounts, and the honest limits](https://everpod.ai/guides/openclaw-google-chat): Google Chat needs a Cloud project, a service account, and a public HTTPS webhook. Its API won't let a bot upload files or react. The full setup, plus what to expect.
- [Connecting OpenClaw to Microsoft Teams, step by step](https://everpod.ai/guides/openclaw-microsoft-teams): Teams ships bundled with OpenClaw, but the platform side runs through Microsoft's preview CLI, an Azure bot, and a tunnel. The working path, the single-tenant rule, and the file-sending catch.
- [Connecting OpenClaw to Matrix (with real end-to-end encryption)](https://everpod.ai/guides/openclaw-matrix): The lightest channel setup OpenClaw has: any homeserver, an access token, no webhook, no public URL, and full E2EE support. Setup, encryption upkeep, and the LAN-homeserver trap.
- [OpenClaw and Zalo: what works today, and which Zalo you mean](https://everpod.ai/guides/openclaw-zalo): The official plugin covers Zalo Bot-platform bots (not Official Accounts, not personal-account automation), and it's experimental. Token setup, long-polling, and the real limits.
- [What is Buzz? Block's open-source workspace where agents are members](https://everpod.ai/guides/what-is-buzz): Buzz launched July 2026: chat, git, and AI agents in one app, on a Nostr relay you can self-host. Every agent signs its own work with its own key. What ships today, and what's still early.
- [OpenClaw's Telegram pairing message, decoded (and how to approve it)](https://everpod.ai/guides/openclaw-telegram-pairing): A first message to a fresh bot gets a pairing code and instructions written for the wrong reader. What the template means, the one command that approves it, and the one-hour expiry.
- [What is a webhook? Explained for agent owners](https://everpod.ai/guides/what-is-a-webhook): An API is a door you knock on; a webhook is the service knocking on yours. How deliveries work, why receivers verify signatures, and which agent channels actually need a public URL.
- [Can you create a Telegram account from a desktop computer?](https://everpod.ai/guides/telegram-account-desktop): No. Telegram's server refuses signup from Desktop and Web apps; a new account needs the official mobile app once. Everything after works from a computer, BotFather and bot tokens included.
- [OpenClaw channels: three ways to connect a chat app](https://everpod.ai/guides/openclaw-channels): Choose a channel, connect it through chat, the control panel or the CLI, and check the difference between configured, connected and allowed to reply.
- [Connect Twilio SMS to OpenClaw: setup, webhooks and trial limits](https://everpod.ai/guides/openclaw-twilio-sms): The SMS plugin, Twilio credentials, exact webhook URL and first-message pairing, plus three setup failures witnessed on OpenClaw 2026.9.1.

## Guides: Running & maintaining

- [Your OpenClaw agent's server stopped on its own? Why, and the fix](https://everpod.ai/guides/openclaw-background-process-stopped): A server your OpenClaw agent starts with & or nohup stops with its command, or at the exec tool's 30-minute limit. What we saw, and how to keep one running.
- [OpenClaw "Sandbox mode requires Docker" error: causes and fixes](https://everpod.ai/guides/openclaw-sandbox-mode-requires-docker): OpenClaw only calls Docker for sessions a setting has sandboxed. Which ones fail, how to find the setting, reach Docker, switch backend or turn it off.
- [AI email assistant: an agent that keeps on top of your inbox](https://everpod.ai/guides/ai-email-assistant): Give your agent a Gmail account of its own and forward it your email. It checks on your schedule, tells you what needs an answer and drafts replies you approve.
- [AI lease review: have your agent check a lease before you sign](https://everpod.ai/guides/ai-lease-review): Send your agent the lease from your phone and get the clauses that could cost you, what each one means, and a reminder before the notice deadline.
- [OpenRouter "requires more credits, or fewer max_tokens" explained](https://everpod.ai/guides/openrouter-requires-more-credits): OpenRouter prices a request's whole possible output before it runs, so an agent stops with money showing. What is stranded per model, and the two fixes.
- [Use OpenClaw as your fitness coach](https://everpod.ai/guides/openclaw-fitness-coach): Give OpenClaw your training plan, report workouts from your phone, and review your progress with a coach that keeps your goals and history.
- [Turn recipe screenshots into a meal planner with OpenClaw](https://everpod.ai/guides/openclaw-meal-planner): Send your saved recipes to OpenClaw and get a planner you can use on your phone, with dinner swaps, portion sizes and one combined shopping list.
- [OpenClaw PDF reading: fix No PDF model configured](https://everpod.ai/guides/openclaw-pdf): Fix OpenClaw PDF-model selection, read text and scanned PDFs, and distinguish a missing model from a file-access problem. Includes a tested configuration fix.
- [OpenClaw request timed out: did the work finish?](https://everpod.ai/guides/openclaw-request-timed-out): Recover completed work after an OpenClaw timeout. Distinguish client waits, agent runtime, CLI deadlines and idle connections before repeating the task.
- [Automate invoice data extraction with a cloud agent](https://everpod.ai/guides/invoice-extraction-cloud-agent): Give your agent an Invoice Data Extraction API key and it takes invoices from a Gmail inbox or your messaging app, keeps your records up to date and reports back.
- [How to connect OpenClaw to Gmail](https://everpod.ai/guides/openclaw-gmail): Read Gmail, process attachments and reply with OpenClaw. A tested email workflow, app-password access, and what to verify before scheduling it.
- [How to share files with OpenClaw](https://everpod.ai/guides/openclaw-share-files): Send documents to your agent and get finished files back. Chat attachments, selected shared folders, and why a server path is not a download.
- [Connect OpenClaw to an MCP server](https://everpod.ai/guides/openclaw-mcp): Add a local or remote MCP server, authorize its account and verify a real tool call. Fix missing tools without confusing setup with a working connection.
- [OpenClaw compaction: when context is full](https://everpod.ai/guides/openclaw-compaction): Use /compact to continue a long OpenClaw conversation. How compaction differs from pruning and /new, and what to check when context still will not fit.
- [How to uninstall OpenClaw completely](https://everpod.ai/guides/how-to-uninstall-openclaw): Removing OpenClaw properly takes three steps: stop the background service, remove the package, and decide what happens to your agent's data. In order.
- [How to update OpenClaw without breaking your agent](https://everpod.ai/guides/how-to-update-openclaw): Native and Docker updates, verified backups, and recovery from a real 2026.9.4 restart loop and memory-index change without discarding agent state.
- [How to move OpenClaw to a new computer without losing your agent](https://everpod.ai/guides/move-openclaw-to-a-new-computer): Move an OpenClaw agent to a new Mac, server or host: the full backup, the restore, what has to be set up again, and why the old Gateway must stay off.
- [OpenClaw update broke your agent? How to roll back or downgrade](https://everpod.ai/guides/openclaw-downgrade): Check whether OpenClaw already rolled back, repair before downgrading, return to a known-good version, and what to do when the old version can't read your data.
- [Backing up an OpenClaw agent (and actually restoring it)](https://everpod.ai/guides/openclaw-backup-and-restore): openclaw backup create and restore: what the archive holds, how a restore goes back into place, scheduled backups, and why every backup carries your agent's secrets.
- [How to fix OpenClaw's “pairing required” (1008) scope error](https://everpod.ai/guides/openclaw-pairing-required): The 1008 close code means a device wants more scopes than an operator approved. Reading Requested vs Approved, approving the request, and the agent-hands case.
- [How to fix “memory search is paused” in OpenClaw](https://everpod.ai/guides/openclaw-memory-search-paused): Vector search pauses when the index no longer matches your settings, or when embeddings quietly fail on subscription-only sign-ins. Both causes, both fixes.
- [What's the latest OpenClaw version, and should you update?](https://everpod.ai/guides/openclaw-latest-version): Whether to update to the latest OpenClaw release, what each recent release fixes and changes, what to do first, and how to check the version your agent runs.
- [How to turn off OpenClaw's update banner (and why Update now does nothing)](https://everpod.ai/guides/openclaw-update-banner): The real key is update.checkOnStart: false. The widely-cited updates.notify never shipped and crash-loops pinned gateways. Plus the not-git-install button, explained.
- [Giving your OpenClaw agent access to a GitHub repo, the safe way](https://everpod.ai/guides/openclaw-github): The agent generates a keypair, you add the public half as a single-repo deploy key, and no token ever touches the chat. Plus when you need a fine-grained PAT instead.
- [How to stop, restart, or disable OpenClaw (without uninstalling)](https://everpod.ai/guides/how-to-restart-openclaw): The gateway stop/start/restart commands, why the process resurrects after you kill it, what actually stops while it's down, and how to make 'off' stick per platform.
- [OpenClaw cron jobs: putting your agent on a schedule](https://everpod.ai/guides/openclaw-cron): The built-in Automations scheduler: reminders, recurring briefings, fire-on-exit triggers. The UTC gotcha, per-run token costs, and why a sleeping laptop skips the night.
- [OpenClaw hooks: run scripts when your agent starts, resets, or speaks](https://everpod.ai/guides/openclaw-hooks): Hooks are event scripts inside the Gateway, distinct from Automations (schedules), skills (instructions), and webhooks (inbound HTTP, confusingly under the same config key). The full map.
- [Can you run multiple OpenClaw agents on one machine?](https://everpod.ai/guides/openclaw-multiple-agents): Yes: one Gateway hosts many agents with separate workspaces, routed by channel bindings. What stays shared, when docs say to run separate gateways instead, and how far the isolation goes.
- [Keeping your agent's Ubuntu server patched, without breaking anything](https://everpod.ai/guides/agent-vps-patching): Ubuntu already auto-installs security updates. The real decisions are reboots and what not to auto-upgrade. How unattended-upgrades works, the reboot-required mechanics, and the agent-box policy.
- [OpenClaw won't start after a config edit: crash loops, the breaker, and the clean fix](https://everpod.ai/guides/openclaw-gateway-crash-loop): The Gateway fails closed on invalid config, and a bad live edit hides until the next restart. The validate-first workflow, doctor's last-known-good restore, and how the crash-loop breaker behaves.
- [Hermes kanban: a task board your agents work through](https://everpod.ai/guides/hermes-kanban): Nous Research's multi-agent feature: named profiles claim tasks from a shared SQLite board, a dispatcher caps concurrency and spend, and /kanban runs it all from Telegram. How it actually works.
- [What is tmux? The tool that keeps agent sessions alive](https://everpod.ai/guides/what-is-tmux): A terminal multiplexer: programs keep running after you disconnect, and you reattach from anywhere. The five commands that matter, tmux vs screen, and the one thing tmux can't survive.
- [Docker Compose .env vs env_file: two different things](https://everpod.ai/guides/docker-compose-env-file): One fills placeholders in your compose file; the other pours variables into the container. Point env_file at your project .env and host paths ride into an environment where they're wrong.
- [Git “detected dubious ownership”: the fix, sized right](https://everpod.ai/guides/git-dubious-ownership): The error means the repo's owner isn't the user running git: a real protection, not a nuisance. The one-line fix, the wildcard to avoid, and when the mismatch is the actual bug.
- [What's new in OpenClaw 2.0, and should you upgrade yet?](https://everpod.ai/guides/whats-new-in-openclaw-2): OpenClaw 2.0 rebuilds the web app, stops resetting conversations daily, adds self-consolidating memory and a model allowlist, and migrates older installs with openclaw doctor --fix. What changes, what breaks, and how to upgrade now that the release-day doctor bug is fixed.
- [What openclaw doctor --fix actually does](https://everpod.ai/guides/openclaw-doctor): Doctor can migrate state even without --fix. The read-only and repair modes, what --fix changes, and the historical headless migration trap.
- [How to change the model in OpenClaw (and why a switch might not stick)](https://everpod.ai/guides/openclaw-change-model): /model with -s, -a, or -g scopes since 2.0, the modelPolicy.allow allowlist that replaced the old models map, and the traps: pending runs, wrong scope, and 7.x's stale-channel capture.
- [Your OpenClaw agent stopped answering? Check the model before the server](https://everpod.ai/guides/openclaw-agent-not-responding): A lapsed ChatGPT subscription silently gates the model your agent is set to: every turn fails while the machine looks healthy. The one-command escape, and the triage order that finds it fast.
- [OpenClaw heartbeat: what your agent does every 30 minutes, and what it costs](https://everpod.ai/guides/openclaw-heartbeat): OpenClaw's heartbeat wakes your agent twice an hour and carries your whole conversation each time: 28k to 107k tokens a run, measured. The cost, and how to cut it.
- [“OpenClaw change was cancelled because its run ended”: the approval card, explained](https://everpod.ai/guides/openclaw-change-cancelled): A cancelled approval could be a bug in OpenClaw 2026.9.1, fixed in 2026.9.2. How to check the version and distinguish the bug from a stopped or expired request.
- [“No usable sandbox!”: why Chromium won't start for your agent in Docker on Ubuntu 24.04](https://everpod.ai/guides/chrome-no-usable-sandbox-docker): Two walls stop an agent's browser in a container: Docker's default seccomp profile refuses the user namespaces Chromium's sandbox needs, and Ubuntu 23.10+ restricts them with AppArmor. The seccomp rule, the sysctl, and why noSandbox is the last resort.
- [“could not read Username for 'https://github.com'” on a public repo: why a cloud server gets asked for credentials](https://everpod.ai/guides/git-could-not-read-username-public-repo): GitHub answers some anonymous protocol-v2 fetches over HTTP/2 from datacenter ranges with a 401, and git turns it into a username prompt plus “expected flush after ref listing”. The trace, the one-line fix, and what breaks downstream.
- [Why an AI agent's browser can kill a small server: Chromium tabs, the OOM killer, and swap](https://everpod.ai/guides/agent-browser-out-of-memory): After ten browsing tasks an agent's Chromium held 31 processes on a 2 vCPU, 4 GB server with no swap; the kernel killed a renderer and every later task stalled. What we measured, why a laptop shrugs it off, and the fixes.
- [OpenClaw mission control: running a squad of agents on a computer of your own](https://everpod.ai/guides/openclaw-mission-control): Mission control is the layer that turns several agents into a team: a shared board, agents that wake each other, schedules, approval gates, a cost ledger. What OpenClaw 2 ships natively, what Squad and Grok Bot charge for, and how to build it yourself.
- [Connect n8n to OpenClaw: trigger an agent through a webhook](https://everpod.ai/guides/openclaw-n8n-webhooks): A working n8n-to-OpenClaw request, the JSON body and token it needs, and why HTTP acceptance is not proof of delivery. The same webhook pattern for Zapier and Make.
- [OpenClaw WhatsApp: "Connection failed" and device-linking errors](https://everpod.ai/guides/openclaw-whatsapp-connection-failed): What failed before WhatsApp linked, what succeeded the next day on the same host and version, and why repeated QR scans are not a diagnosis.
- [OpenClaw: "question answer caller policy does not match its creator"](https://everpod.ai/guides/openclaw-question-answer-policy): Recover from a rejected answer or a Telegram question that never arrives. What the 2026.9.4 reports establish, and the status of the proposed delivery fix.
- [OpenClaw: "Managed local embeddings are unavailable"](https://everpod.ai/guides/openclaw-managed-local-embeddings-unavailable): When a saved llama.cpp server path points to a missing executable, reindexing cannot fix it. Restore the provider on the Gateway host, then rebuild memory search.

## Guides: Hosting choices

- [OpenAI dots vs Meta Muse vs OpenClaw: a lab's agent or yours](https://everpod.ai/guides/openai-dots-vs-meta-muse-vs-openclaw): Dots and Muse are agents OpenAI and Meta run for you on their cloud; OpenClaw is yours, on a computer you control. Compared on price, models, apps and memory.
- [Mac VPS for AI agents: renting a Mac mini in the cloud](https://everpod.ai/guides/mac-vps-for-ai-agents): Rent a Mac mini in the cloud from about $70 a month, or hourly on AWS. What each option costs, what Apple's license allows, and when an agent needs macOS.
- [Codex CLI on a Linux server: install, login, bubblewrap](https://everpod.ai/guides/codex-cli-linux-server): Install Codex CLI on a VPS, sign in with codex login --device-auth, get its bubblewrap sandbox working on Ubuntu 24.04, and fix command not found over SSH.
- [Claude setup-token: the one-year token and what it can't do](https://everpod.ai/guides/claude-setup-token): claude setup-token prints a one-year Claude subscription token for scripts and CI. It can't run Remote Control, so a lapsed login needs /login on the machine.
- [The best VPS for OpenClaw: real requirements, not guesswork](https://everpod.ai/guides/best-vps-for-openclaw): What OpenClaw actually needs from a server, measured rather than assumed, and how to choose a provider, size, and setup that won't bite you later.
- [OpenClaw alternatives in 2026: other agents, and other ways to run it](https://everpod.ai/guides/openclaw-alternatives): Searching for an OpenClaw alternative usually means one of two things: a different agent, or a less painful way to run this one. Both, compared.
- [OpenClaw vs Hermes: two philosophies of open-source agent](https://everpod.ai/guides/openclaw-vs-hermes): The two most-watched open-source agents of 2026 solve different problems. How they differ on architecture, skills, memory, and security, and how to choose.
- [Should you run OpenClaw locally or on a VPS?](https://everpod.ai/guides/run-openclaw-locally-or-on-a-vps): Your laptop is fine for trying OpenClaw out, and wrong for living with it. How to decide where your agent should actually run.
- [Managed OpenClaw hosting vs self-hosting on a VPS](https://everpod.ai/guides/managed-openclaw-hosting-vs-self-hosting): Renting a server and doing it yourself versus paying someone to run OpenClaw for you: the work, the risks, and the costs, compared.
- [KiloClaw is shutting down: how to save your agent and where to move it](https://everpod.ai/guides/kiloclaw-shutting-down): KiloClaw retires on 20 October 2026. What the dashboard export saves, the fuller backup to ask your agent for, what to reconnect, and the alternatives.
- [Hostinger OpenClaw, managed or VPS: what you get and what you pay](https://everpod.ai/guides/hostinger-managed-openclaw): Hostinger's Managed OpenClaw and OpenClaw VPS told apart: prices per term, the model credits, setup, updates, backups, and the dashboard you manage it in.
- [Managed OpenClaw hosting compared: Hostinger, MyClaw, Cloudways and more](https://everpod.ai/guides/managed-openclaw-hosting-compared): Nine OpenClaw hosts side by side from their own pages: price and term, model usage, updates, backups, bringing your agent in, and what you still do yourself.
- [Mac Mini vs VPS vs managed pod: where should your OpenClaw agent live?](https://everpod.ai/guides/mac-mini-vs-vps-vs-managed-pod): Compare a Mac Mini, a self-managed VPS and an Everpod pod for OpenClaw: local access, model costs, backups and who maintains the computer.
- [OpenClaw vs Claude Cowork: an agent you own, or a product you rent?](https://everpod.ai/guides/openclaw-vs-claude-cowork): Anthropic's polished in-app agent against the open-source agent that lives on your machine, compared on models, cost shape, data control, and vendor risk.
- [Running Claude Code on an always-on VPS: the tmux pattern, done right](https://everpod.ai/guides/claude-code-on-a-vps): The persistent coding-agent box developers keep rebuilding: tmux for sessions that survive, a handover CLAUDE.md, and the security posture that makes it sane.
- [OpenClaw on Hetzner: the setup guide we actually operate by](https://everpod.ai/guides/openclaw-hetzner): Server sizing from measured usage, the official Docker path, and the hardening the official guide skips, from a team running production OpenClaw pods on Hetzner.
- [Claude Cowork alternatives: what to pick, by what you're missing](https://everpod.ai/guides/claude-cowork-alternatives): Ownership, model choice, usage limits, or just the coding half: the map from Cowork to OpenClaw, Hermes, Claude Code/Codex, and the DIY routes.
- [OpenClaw vs Claude Code: which agent do you actually want?](https://everpod.ai/guides/openclaw-vs-claude-code): Anthropic's coding agent against the open-source personal agent, on models, cost, channels, and ownership. And the stacking trick: OpenClaw can drive Claude Code as a backend.
- [How to use Claude Code from your phone](https://everpod.ai/guides/claude-code-from-your-phone): The Claude app now runs Code on your phone: cloud sessions, Remote Control into your own machine, and chat channels. What runs where, which plans get what, and the one thing every path needs.
- [What is Forgejo? The community-owned Git forge, explained](https://everpod.ai/guides/what-is-forgejo): A self-hosted GitHub alternative forked from Gitea in 2022, GPL-licensed and run under a non-profit, and lately the git server people give their AI agents. Origins, vs-Gitea, and setup shapes.
- [Local AI box vs cloud: should you buy your agent a computer?](https://everpod.ai/guides/local-ai-box-vs-cloud): Local AI hardware is really a memory purchase, and memory prices surged in 2026, with vendors saying so openly. The ladder from $249 to $6,950, the math against subscriptions, and when local wins.
- [Codex remote control: driving your dev machine from your phone](https://everpod.ai/guides/codex-remote-control): Codex Remote pairs the ChatGPT phone app with a Mac or Windows host by QR code. Setup, plans, codex remote-control, and what happens when the host sleeps.
- [What is a VPS? A plain-language guide for agent owners](https://everpod.ai/guides/what-is-a-vps): A virtual private server is a rented slice of a real computer that stays on when your laptop doesn't. What you actually get for a few euros a month, and why AI agents changed who needs one.
- [Grok Bot vs OpenClaw: rented AI teammates, or an agent you own?](https://everpod.ai/guides/grok-bot-vs-openclaw): xAI's polished bot roster against the open-source agent that lives on your machine, compared on the computer underneath, models, channels, cost shape, and lock-in.
- [Grok Bot alternatives: the same always-on agents, on your terms](https://everpod.ai/guides/grok-bot-alternatives): Want other models, uneasy about every bot sharing one computer and its logins, or not keen on a Cursor or SuperGrok subscription as the price of admission? The open-source routes to the same job, compared.
- [OpenClaw 2.0 cloud sessions and nodes: do you still need a server?](https://everpod.ai/guides/openclaw-cloud-sessions): Sessions can now run on the Gateway, your own paired hardware, or throwaway cloud workers leased through Crabbox. What each piece does, and why the always-on Gateway didn't go anywhere.
- [The best VM or sandbox for AI agents, by the job you're hiring it for](https://everpod.ai/guides/best-vm-for-ai-agents): One persistent agent wants a VPS or managed pod; parallel disposable runs want a sandbox API like Modal or Daytona; teams want a workspace layer. The map, including the ephemeral tax.
- [What is a cloud agent? An always-on AI agent with its own computer, explained](https://everpod.ai/guides/what-is-a-cloud-agent): A cloud agent is an AI agent that runs on a computer of its own in the cloud: always on, with a browser, files, logins and memory that persist. Where the term came from, the two shapes that share it, and the three ways people get one.
- [Why does an AI agent need its own computer?](https://everpod.ai/guides/why-an-ai-agent-needs-its-own-computer): A chat window forgets and a laptop sleeps. An agent that keeps logins, files, a browser and a schedule needs a machine that is always there and nobody else uses: the four things a computer gives it, and why a shared one is the wrong shape.
- [OpenClaw vs Hermes vs Grok Bot: what's actually different, and where each one runs](https://everpod.ai/guides/openclaw-vs-hermes-vs-grok-bot): Two open-source harnesses you host and one closed product on xAI's cloud. Licence, where the machine is, what persists, model choice, memory, channels and price in one table, with the hosting column the popular comparisons skip.
- [What AI teammates actually cost per month: Grok Bot, Squad, Hermes Cloud, and running your own](https://everpod.ai/guides/ai-teammates-cost): The sticker is never the total. Grok Bot rides a $20 to $300 subscription plus metered usage; Squad is $99 plus the ChatGPT plan it recommends; a self-hosted squad is a small server plus the model bill; a managed pod is $29 with usage included.
- [Hermes agent browser on a server: can it use your Chrome logins?](https://everpod.ai/guides/hermes-browser-on-a-server): Hermes' real-profile browsing copies the Chrome profile on the machine Hermes runs on, so a Hermes on a VPS has no logins of yours to borrow and runs headless besides. What the feature does, why it is same-machine by design, and the options for a hosted Hermes.

## Guides: Security

- [Tailscale SSH for an AI agent server: tags, ACLs, key expiry](https://everpod.ai/guides/tailscale-ssh-ai-agent-server): By default an AI agent's server on your tailnet can reach your laptop. Tag it, let only your devices connect, SSH in with no keys, and turn off key expiry.
- [Tailscale and NordVPN together: why it breaks and what works](https://everpod.ai/guides/tailscale-nordvpn): With NordVPN on, tailscale ping answers but ping and SSH to your tailnet fail. Why NordVPN blocks Tailscale, and what works, Mullvad exit nodes included.
- [Is DeepSeek safe for your AI agent? Where your prompts actually go](https://everpod.ai/guides/is-deepseek-safe-for-ai-agents): DeepSeek's weights are open, so the host decides where prompts go. Which of its 25 hosts train, keep data or sit in China, and the two settings that decide it.
- [OpenClaw secrets: Protected secret or Agent-readable environment?](https://everpod.ai/guides/openclaw-protected-secrets): Choose the right access mode for an OpenClaw credential. Why a saved secret may not reach a script, what oc-sent values mean, and how to test without printing keys.
- [OpenClaw pairing, allowlists, and dmPolicy, explained](https://everpod.ai/guides/openclaw-pairing-and-allowlists): Who can talk to your agent is a setting, and the default is nobody you haven't approved. How pairing codes, allowlists, and group mention-gating work.
- [Is it safe to run OpenClaw on your personal computer?](https://everpod.ai/guides/is-openclaw-safe-on-your-personal-computer): OpenClaw is personal-by-default and binds to your machine only. But an autonomous agent on your laptop makes your laptop the security boundary. What that means.
- [Why you should never put your OpenClaw Gateway on the public internet](https://everpod.ai/guides/never-expose-your-openclaw-gateway): OpenClaw's own docs say never expose the Gateway unauthenticated. What can go wrong, and the right ways to reach your agent remotely.
- [What a safe OpenClaw hosting setup should include](https://everpod.ai/guides/safe-openclaw-hosting-checklist): A plain-language checklist for hosting an autonomous agent safely: private gateway, one agent per machine, pairing, pinned versions, backups, and audits.
- [How to run OpenClaw's security audit (and read the results)](https://everpod.ai/guides/openclaw-security-audit): OpenClaw ships a built-in security audit. When to run it, what --deep and --fix do, and how to make sense of the findings.
- [Why your VPS logs show SSH login attempts (and when to worry)](https://everpod.ai/guides/vps-ssh-login-attempts): Fresh servers get automated login attempts within hours, thousands per week. Why it's background noise with keys-only auth, and the three patterns that aren't.
- [SSH says “REMOTE HOST IDENTIFICATION HAS CHANGED” — how to fix it safely](https://everpod.ai/guides/ssh-remote-host-identification-changed): Usually it means the server was rebuilt and the fix is ssh-keygen -R. When it isn't, it matters. Both cases, plus the accept-new pattern for provisioning scripts.
- [Tailscale vs Cloudflare Tunnel for reaching a private agent server](https://everpod.ai/guides/tailscale-vs-cloudflare-tunnel): Private mesh with a client on every device, or public URL with auth at the edge. Opposite shapes, both zero inbound ports. The fork: who needs to connect?
- [OpenClaw sandbox: modes, Docker, Podman and exec approvals](https://everpod.ai/guides/openclaw-sandbox): Off by default, so exec runs on the host. Sandbox modes, the Docker, Podman, SSH, OpenShell and Crabbox backends, what a sandbox can reach, and exec approvals.
- [What is prompt injection? A plain guide for people who run AI agents](https://everpod.ai/guides/what-is-prompt-injection): Text your agent reads can carry instructions it will follow, and no model reliably tells data from commands. Direct vs indirect injection, the lethal trifecta, and the defenses that actually hold.
- [What is mosh? The remote shell that survives roaming, sleep, and bad Wi-Fi](https://everpod.ai/guides/what-is-mosh): Mosh replaces interactive SSH with a UDP session that follows you across networks and laptop sleeps. How it works, the scrollback and forwarding trade-offs, and the tmux pairing.
- [SSH from your iPhone: the toolkit for reaching your server from anywhere](https://everpod.ai/guides/ssh-from-your-phone): iOS kills background connections by design, so the working setup is an SSH client plus mosh plus tmux. Termius, Blink, and Prompt compared, and what that means for agent servers.
- [“ssh.service is not active, cannot reload” — Ubuntu's socket-activated SSH, explained](https://everpod.ai/guides/ssh-service-not-active): Since 22.10, sshd doesn't run until someone connects, so reload fails and set -e scripts die. Why Ubuntu changed it, the commands that work now, and the official way back.
- [Who can see your OpenClaw agent's conversations?](https://everpod.ai/guides/who-can-see-openclaw-conversations): OpenClaw stores everything locally, but every message still passes through a chat platform and a model provider. Who can technically read what, and the dials that change it.
- [How to check open ports on a Linux server — from inside and outside](https://everpod.ai/guides/check-open-ports-linux): ss -tulpn shows what's listening; nmap from another machine shows what the internet sees; and Docker-published ports bypass your firewall rules. The audit, plus what an agent box should show.
- [How to disable SSH password login on Ubuntu — without locking yourself out](https://everpod.ai/guides/disable-ssh-password-login): Keys proven first, then a one-file drop-in named to sort before cloud-init's, because in sshd_config the first value wins. The sshd -t check and the second-session test that make it safe.
- [Should OpenClaw run as root on your server?](https://everpod.ai/guides/openclaw-run-as-root): No, and the official Docker image already refuses to: it runs as node, uid 1000. Why an agent that executes commands deserves its own locked-down user, and the setup that does it properly.
- [What is SSH? A plain-language guide to Secure Shell](https://everpod.ai/guides/what-is-ssh): The encrypted remote-login protocol behind nearly all server administration, your agent's included. Keys vs passwords, how host fingerprints work, and the warnings you'll actually meet.
- [What is a reverse proxy? (And when login belongs in front)](https://everpod.ai/guides/what-is-a-reverse-proxy): A server that faces the internet so your real servers don't: TLS, load balancing, caching, and in identity-aware setups, authentication itself. How the headers work and what makes it safe.
- [Does OpenRouter train on your prompts? Data policy, explained](https://everpod.ai/guides/openrouter-privacy): OpenRouter itself retains nothing unless you opt into logging. Training risk lives with the providers it routes to. The data_collection and zdr controls, what they cover, and what they cost you.
- [Is it safe to run your AI agent on a cloud machine, or should it stay on one you can unplug?](https://everpod.ai/guides/is-it-safe-to-run-an-ai-agent-in-the-cloud): A security writer's case for a Mac mini at home, taken seriously: what a host can and cannot see, what “unplug” maps to on a rented machine, where local genuinely wins, and Everpod's own answer.
- [OpenClaw SecretRefs: fix the plaintext-secret config warning](https://everpod.ai/guides/openclaw-secretrefs): Move supported credentials into OpenClaw's secret store, reference them from configuration and verify the result. What SecretRefs protect, and what they do not.

## Guides: Agents on the web

- [What is AG-UI? The Agent-User Interaction protocol](https://everpod.ai/guides/what-is-ag-ui): AG-UI streams an agent's work to an app as typed events over HTTP. What 1.0 changed, how it differs from MCP, A2A and ACP, and whether OpenClaw speaks it.
- [Can your AI agent post videos to TikTok and YouTube for you?](https://everpod.ai/guides/ai-agent-post-videos-tiktok-youtube): Both posting APIs are free but audited: until then, YouTube locks uploads private and TikTok shows posts only to you. The two routes left, and what they cost.
- [What “[email protected]” means, and how to find the real address](https://everpod.ai/guides/email-protected-cloudflare): Cloudflare swaps each email address on a page for this text until a browser runs its script. How to find the real address, and what AI assistants see instead.
- [Why websites block AI agents — and what an operator can do about it](https://everpod.ai/guides/why-websites-block-ai-agents): Bot walls can't tell your personal agent from a scraper, and there's no passlist for it. How the defenses work, what's legitimate, and what to skip.
- [Cloudflare Bot Fight Mode and AI agents: what it blocks, what it can't tell apart](https://everpod.ai/guides/cloudflare-bot-fight-mode-ai-agents): The free all-or-nothing bot defense challenges AI assistants, APIs, and scrapers alike. What it does, where the granular AI controls actually live, and both sides' playbook.
- [Should you block AI crawlers? Training, search, and agent bots, separated](https://everpod.ai/guides/should-you-block-ai-crawlers): GPTBot and ClaudeBot train models; OAI-SearchBot surfaces you in AI search; ChatGPT-User fetches for a live human. The robots.txt policy that says what you mean.
- [What is llms.txt — and should your site have one?](https://everpod.ai/guides/what-is-llms-txt): A proposed markdown index for AI systems at /llms.txt. What the spec says, who actually reads the file (fewer than you've heard), and where it genuinely helps today.
- [OpenClaw browser and Chrome extension: how to set them up](https://everpod.ai/guides/openclaw-browser): The agent's own isolated browser, or your signed-in Chrome through remote debugging or the OpenClaw Chrome extension: what each can reach and how to set it up.
- [What is a HAR file? Capture, risks, and what an AI agent can do with one](https://everpod.ai/guides/what-is-a-har-file): A HAR is a JSON recording of every request a page made: headers, cookies, bodies, timings. How to export one in each browser, why support teams ask for them, and the token-theft caution.
- [How to make your website readable by AI agents](https://everpod.ai/guides/make-your-website-agent-readable): Not all bots run JavaScript, and agents buy things now. The build-side checklist: server-rendered substance, a robots.txt that says what you mean, JSON-LD, markdown twins, and machine-readable pricing.
- [How to write an llms.txt for your site (with a copy-paste template)](https://everpod.ai/guides/create-an-llms-txt): The spec is one H1, a blockquote, and H2 link lists: five minutes of work. The exact format, a template, three live examples, and what llms-full.txt actually is (not the spec).
- [What is the Agent Client Protocol (ACP)?](https://everpod.ai/guides/what-is-acp): Zed's open standard lets any coding agent talk to any editor: JSON-RPC over stdio, reusing MCP's JSON types. Who supports it, how it differs from MCP, and which of the three “ACPs” you've met.
- [What is MCP? The Model Context Protocol, explained](https://everpod.ai/guides/what-is-mcp): An open standard connecting AI apps to tools, data, and prompts: a USB-C port for AI, in its own docs' words. Where it came from, how it works, and how it differs from an API.
- [What is an MCP server, and should you run one?](https://everpod.ai/guides/what-is-an-mcp-server): A small program that gives AI applications tools, resources, and prompts over the Model Context Protocol. Local vs remote, the official registry, and the trust rules before you connect one.
- [What is Kitesurf? Cloudflare's browser for AI agents](https://everpod.ai/guides/what-is-kitesurf): A Rust rendering engine on Cloudflare Workers, no Chromium, announced August 2026 and free in beta. What it renders, the benchmark trade-offs, and how agents drive it.
- [Which websites block an AI agent running on a cloud server? We measured 83 URLs](https://everpod.ai/guides/which-websites-block-ai-agents-on-a-server): An 83-URL test comparing headless and headed Chromium on a cloud server with a laptop via VPN. Site-by-site results, test conditions and what they establish.
- [Tailscale exit node for a cloud agent: browse from home, run in the cloud](https://everpod.ai/guides/tailscale-exit-node): An exit node routes a device's internet traffic out through another device on your Tailscale network. Put one on a machine at home and a cloud server's browser leaves from your household address, which is what the sites that block datacenter ranges key on.