llms.txt
# Documentation
- [What is Buda](/docs): Buda is a cloud-native AI agent workspace — recruit, run, and scale teams of agents that actually do the work.
- **Learn**
- Get Started
- [Quickstart — hire your first agent](/docs/getting-started/quickstart): Go from sign-up to your first deliverable in about five minutes — recruit an agent, give it a task in plain language, and review the result.
- [Run and review your first task](/docs/getting-started/first-task): Write a clear prompt, follow each step across files, browser, terminal and Git, handle waiting-for-input, and iterate to a result you're happy with.
- [Create an account](/docs/getting-started/create-account): Sign up with email, GitHub, or Google, verify your address, and land in your first workspace — no install required.
- [The Buda model](/docs/getting-started/the-buda-model): How the CEO → Organizer → Claws model maps to Spaces, agents, and the cloud computer — so you run agents the way a CEO runs a company.
- Core Concepts
- [How Buda works](/docs/concepts/how-buda-works): The end-to-end loop — you direct, the Organizer coordinates, and Claws execute in cloud computers. Read this first.
- [Agents and Claws](/docs/concepts/agents-and-claws): What an agent is, what a Claw executes, and how identity, instructions, Drive, Skills, and channels come together into one AI worker.
- [The cloud computer](/docs/concepts/the-cloud-computer): Every agent gets its own isolated, long-running sandbox — files, browser, terminal, and Git — that persists between tasks. No laptop required.
- [Sessions and tasks](/docs/concepts/sessions-and-tasks): How a session isolates context for one job, how a task moves through its lifecycle, and what happens when an agent is waiting for your input.
- [Spaces, teams and members](/docs/concepts/spaces-teams-members): How User → Space → Team → Agent structures your AI company, and what's shared inside a Space.
- [Drive & Memory](/docs/concepts/drive-and-memory): An agent's memory is just files — and they live in its Drive. Feed it your materials and it stops starting from zero: it remembers you, and gets more like you over time.
- [Skills and automations](/docs/concepts/skills-and-automations): Turn repeatable work into reusable SOPs your agents run on demand or on a schedule — and understand where Skills live: system, agent, or space.
- [Credits and the billing model](/docs/concepts/credits-and-billing-model): How daily, monthly, and balance credit pools work, why they're shared per Space, and how usage is metered as your agents run.
- **Build**
- The Agent Workspace
- [The Agent Workspace](/docs/agent-workspace): Watch your Agent work — chat, files, terminal, browser, preview, IDE, and Git — live in one multi-tab workbench.
- [Agent tab](/docs/agent-workspace/agent-tab): Chat with your Agent, run tasks, switch sessions, and watch every tool call and artifact stream in live.
- [Files (Drive)](/docs/agent-workspace/drive-tab): Browse, upload, and edit your Agent's persistent Drive — the working directory and knowledge base it reads and writes during every task.
- [Terminal](/docs/agent-workspace/terminal-tab): Open a live shell inside your Agent's sandbox and run commands directly from the browser.
- [Browser tabs](/docs/agent-workspace/browser-tab): Control a real browser from inside Buda — AI Browser for an Agent-operated cloud Chromium, Local Browser for localhost and internal tools.
- [Web Preview](/docs/agent-workspace/web-preview): See a web app the Agent is running inside its sandbox — Buda turns its localhost link into a reachable preview URL.
- [VS Code IDE](/docs/agent-workspace/vscode-ide): Connect your local VS Code to an Agent's cloud computer over Remote SSH and edit the workspace from your own IDE.
- [Git](/docs/agent-workspace/git-tab): Review the Agent's work as commits and diffs, switch branches, and roll back safely — a visual audit trail of every code change.
- [High-performance volume](/docs/agent-workspace/high-perf-volume): Choose fast block storage for code-heavy Agents — installs, builds, and Git run dramatically faster than on standard storage.
- [Layouts](/docs/agent-workspace/layouts): Switch between Company Space Layout (manage a whole Space of agents) and Single Agent Layout (focus on one Agent's sessions).
- [Projects and folders](/docs/agent-workspace/projects-and-folders): Keep a growing roster of agents tidy with folders, and group an Agent's related sessions into projects.
- [Environment variables](/docs/agent-workspace/environment-variables): Give an Agent the API keys and config its sandbox needs — set per-Agent environment variables once and they're available everywhere it runs.
- [Runtime monitoring](/docs/agent-workspace/runtime-monitoring): See your Agent's open ports, CPU and memory usage, and sandbox status — set its sleep timeout and restart it when needed.
- Working with Agents
- [Create an agent](/docs/guides/create-an-agent): Hire a new agent — name it, give it an icon and skills — then put it to work in its own cloud computer.
- [Run and review tasks](/docs/guides/run-and-review-tasks): Give an agent a task in plain language, watch it work step by step, answer when it needs you, and collect the results.
- [Scheduled tasks](/docs/guides/scheduled-tasks): Run an agent task automatically — on a cron, a fixed interval, or once at a future time — and review every run.
- [Share and collaborate](/docs/guides/share-and-collaborate): Show off a session with a public link, or invite teammates and teams to an agent as viewers, editors, or owners.
- [Agent transfer](/docs/guides/agent-transfer): Move an agent — with all its sessions, files, and skills — permanently from one Space to another.
- [Sessions dashboard](/docs/guides/sessions-dashboard): One place to see every session across your whole workspace — and instantly spot the ones waiting on you.
- [AI Workforce Commander](/docs/guides/commander): Your operations console — every session, every running app, and every deployed site for the whole workforce, in one place.
- [Email to an agent](/docs/guides/email-to-agent): Give an agent its own email address so inbound mail becomes a task it works on — and replies to — in the same thread.
- [Deploy and publish sites](/docs/guides/deploy-sites): Take a site an agent built in Drive, preview it, publish it to a live public URL — and unpublish whenever you want.
- [Use a custom domain](/docs/guides/custom-domains): Point an already published Buda Site to a domain you own.
- [Preview web apps](/docs/guides/preview-apps): See an app your agent is building, instantly — a live, throwaway preview of its dev server, no deploy required.
- Skills & Automations
- [Skills overview](/docs/skills): Capabilities you install into agents — built-in, custom, or from the marketplace.
- [Install and manage skills](/docs/skills/install-and-manage): Browse, scope, refresh, and read the Skills available to your agents.
- [Built-in skills](/docs/skills/built-in): FAR, Find Skills, Skill Creator, Buda Automations, App Builder, Web Preview, and Machine Info.
- [Create a custom skill](/docs/skills/create-custom): Build, test, and ship your own skill with the Skill Creator.
- [SKILL.md format](/docs/skills/skill-manifest): The skill manifest: structure, required fields, tools, inputs, and examples.
- [Automations](/docs/skills/automations): Schedule skills and tasks as recurring SOPs your agents run on their own.
- [Video generation](/docs/skills/video-generation): Turn text or reference media into videos with the built-in generate_video tool.
- **Connect**
- Channels
- [Channels Overview](/docs/channels): Let your agent answer in WhatsApp, Telegram, Slack, Feishu, and more — channels route inbound messages to an agent and stream replies back.
- [Connect WhatsApp](/docs/channels/whatsapp): Link a WhatsApp account to a Buda agent by scanning a QR code — no Meta Business account or API approval required.
- [Connect WeChat](/docs/channels/wechat): Link a personal WeChat account to a Buda agent by scanning a QR code — no app registration or credentials required.
- [Connect Telegram](/docs/channels/telegram): Create a Telegram bot with BotFather and connect it to a Buda agent for two-way AI conversations in minutes.
- [Connect Discord](/docs/channels/discord): Connect a Discord bot to a Buda agent over the native Gateway — no webhook URL or public endpoint required.
- [Connect Slack](/docs/channels/slack): Connect a Slack bot to a Buda agent with Socket Mode — chat in channels or DMs, no public webhook to host.
- [Connect WeCom](/docs/channels/wecom): Create a WeCom Smart Bot and connect it to a Buda agent so employees can chat with it directly in WeCom.
- [Connect DingTalk](/docs/channels/dingtalk): Connect a DingTalk (钉钉) robot to a Buda agent over Stream Mode — scan to auto-register an app, or enter a Client ID and Secret manually.
- [Connect Feishu / Lark](/docs/channels/feishu): Connect a Feishu (Lark) bot to a Buda agent over a persistent WebSocket — App ID and Secret only, no public webhook.
- [Connect Microsoft Teams](/docs/channels/msteams): Register an Azure Bot, install a Teams app, and connect it to a Buda agent for two-way AI conversations.
- [Embed a web chat widget](/docs/channels/web-widget): Put a public AI chat widget on your own website — create a web channel, embed one script, and choose session, login, and read-only behavior.
- Integrations & Connectors
- [Integrations](/docs/integrations): Give each agent the exact tools and external services it needs — toggle tools per agent and connect accounts over OAuth.
- [OAuth app integrations](/docs/integrations/oauth-apps): Connect the work apps each Buda agent needs, review permissions before approval, and keep every connection under your control.
- [Linear](/docs/integrations/linear): Connect Linear so an agent can read your issues and act on them — triage, comment, and update without you switching apps.
- [Slack (tools)](/docs/integrations/slack-tools): Give an agent Slack access so it can read channels and post updates — turning Slack into a tool the agent uses, not a tab you babysit.
- [Use Agent Connectors](/docs/integrations/connectors): Connect a computer and select it as the runtime device when creating an Agent.
- [MCP with Claude](/docs/integrations/mcp): Drive your Buda agents from Claude Desktop over the Model Context Protocol — authenticate with an API key and call Buda right from your chat.
- Marketplace
- [Marketplace](/docs/marketplace): Browse and install Skills, hire pre-built Agents, and recruit whole Teams — ready-made and live in minutes.
- [Install skills](/docs/marketplace/install-skills): Find free or paid skills and add them to any agent in a few clicks.
- [Hire agents](/docs/marketplace/hire-agents): Bring a pre-configured agent — already loaded with the right skills — into a Space.
- [Recruit teams](/docs/marketplace/recruit-teams): Stand up a coordinated team of agents — each with its own role — in a single step.
- [Sell Your Skills on Buda Marketplace](/docs/marketplace/sell-skills): How to publish and monetize your AI Skills, Agents, and Teams — including private repositories
- [Create a Skill / Agent / Team Repository](/docs/marketplace/create-skill-repo): How to structure a GitHub repository so Buda can discover and publish your Skills, Agents, and Teams
- [Pricing and monetization](/docs/marketplace/pricing-and-monetization): How free, one-time, and subscription pricing work — plus private-repo selling, payouts, and listing review.
- **Manage**
- Spaces, Teams & Members
- [Spaces](/docs/spaces): Create and manage workspaces — each Space is its own AI company, with its own agents, members, and billing.
- [Switch between Spaces](/docs/spaces/switch-spaces): Move between the workspaces you belong to, set where you land, and keep each Space's setup separate.
- [Invite members to a Space](/docs/spaces/invite-members): Invite people into a Space by email and manage pending invitations — resend, cancel, and track who has joined.
- [Roles and permissions](/docs/spaces/roles-and-permissions): Owner, admin, member, and viewer roles in a Space — what each can do, how to change a role, and how to remove or leave.
- [Agent visibility & sharing](/docs/spaces/agent-visibility): Control which agents members can see in a Space — restrict visibility, share an agent with specific people or the whole Space, and manage it all from one place.
- [Teams](/docs/spaces/teams): Group members into teams within a Space — create a team, add or remove people, and delete it when you're done.
- [Labs (experimental features)](/docs/spaces/labs): Opt in to experimental, Space-level features in Buda — what they are, how to enable them, and what to expect.
- Billing & Credits
- [Plans & pricing](/docs/billing): How Buda bills per Space — plans, the three-pool credit system, storage, and what each tier includes.
- [What are credits](/docs/billing/credits): How Buda's daily, monthly, and balance credit pools work — what consumes them, what doesn't, and how to stretch them.
- [Track usage](/docs/billing/usage): Monitor your Space's credit pools, storage, and quotas in real time — and see which models cost the most.
- [Payment methods](/docs/billing/payment-methods): Pay for Buda by card, Alipay, WeChat Pay, or PayPal — plus how invoices and receipts work.
- [Redeem a reward code](/docs/billing/redeem): Turn a Buda reward code into balance credits and recharge them into your Space.
- [Space Drive](/docs/billing/space-drive): The shared /space directory available to all Agents in your Space — your team's persistent shared memory.
- [Referral program](/docs/billing/referral): Share your referral code, earn balance credits when friends join, and track your rewards.
- [Grant programs](/docs/billing/programs): Free time-limited Plus access for students, startups, nonprofits, and open-source maintainers.
- Account
- [Profile and preferences](/docs/account/profile-and-preferences): Set your name and avatar, switch theme and language, tune email notifications, and learn the keyboard shortcuts.
- [Notifications](/docs/account/notifications): How Buda's in-app notification inbox works — where alerts appear, how to clear them, and what triggers them.
- [Delete your account](/docs/account/delete-account): How to permanently close your Buda account, what happens to your data, and how to get back in if you change your mind.
- **Develop**
- Developers
- [Developers (API)](/docs/developers): Run hosted Buda agents from your own code — REST API, embeds, MCP, and a GitHub-connected developer portal.
- [Authentication](/docs/developers/authentication): Create an sk_ API key, authenticate every request, and rotate or expire keys without breaking production.
- [REST API](/docs/developers/rest-api): Buda's contract-first REST API at /api/v1 — OpenAPI spec, Swagger UI, and a worked quick start.
- [Model API](/docs/developers/model-api): Point any OpenAI Responses API client — including OpenAI Codex CLI — at Buda and call any of Buda's models with your existing sk_ API key.
- [API Claws](/docs/developers/api-claws): Run hosted Buda agents programmatically — give your hardware or software product cloud AI agents with no inference infrastructure to operate.
- [Embed agents](/docs/developers/embed): Put a live Buda agent in front of your users — hosted iframe URLs or native embed tokens, without ever shipping your API key.
- [GitHub app](/docs/developers/github-app): Connect your repositories to Buda's Developer Portal to import skills and agents and keep Marketplace listings in sync.
- [MCP reference](/docs/developers/mcp): Buda's Model Context Protocol endpoint — URL, authentication headers, and the tools generated from the REST API.
- [Connect Buda over ACP](/docs/developers/acp): Connect Zed, JetBrains, or your own program to a Buda-hosted agent, step by step.
- OpenAPI Reference
- [OpenAPI Reference](/docs/api-openapi): Browse the complete Buda REST API, generated directly from the live oRPC contracts.
- Api claw
- [Delete API Agent channel](/docs/api-openapi/api-claw/delete-channelId)
- [Cancel API Agent session run](/docs/api-openapi/api-claw/delete-run)
- [Delete API Agent session](/docs/api-openapi/api-claw/delete-sessionId)
- [Delete API Agent scheduled task](/docs/api-openapi/api-claw/delete-taskId)
- [Abort API Agent Drive upload](/docs/api-openapi/api-claw/delete-uploadSessionId)
- [List Claw Agents](/docs/api-openapi/api-claw/get-agents)
- [List API Agent Drive files](/docs/api-openapi/api-claw/get-api-agents-agentId-drive-files)
- [List API Agents](/docs/api-openapi/api-claw/get-api-agents)
- [List API Agent channels](/docs/api-openapi/api-claw/get-channels)
- [List API Agent Drive items](/docs/api-openapi/api-claw/get-items)
- [Get API Agent automation limits](/docs/api-openapi/api-claw/get-limits)
- [List API Agent scheduled task runs](/docs/api-openapi/api-claw/get-runs)
- [List API Agent scheduled tasks](/docs/api-openapi/api-claw/get-scheduled-tasks)
- [Get API Agent session](/docs/api-openapi/api-claw/get-sessionId)
- [List API Agent sessions](/docs/api-openapi/api-claw/get-sessions)
- [List Agent Drive files](/docs/api-openapi/api-claw/get-spaces-spaceId-agents-agentId-drive-files)
- [List Spaces](/docs/api-openapi/api-claw/get-spaces)
- [Resume API Agent session stream](/docs/api-openapi/api-claw/get-stream): Resumes an active AI SDK UI message stream for an API Agent session.
- [Get API Agent scheduled task](/docs/api-openapi/api-claw/get-taskId)
- [Get API Agent Drive upload status](/docs/api-openapi/api-claw/get-uploadSessionId)
- [Update API Agent](/docs/api-openapi/api-claw/patch-agentId)
- [Update API Agent channel](/docs/api-openapi/api-claw/patch-channelId)
- [Rename API Agent session](/docs/api-openapi/api-claw/patch-sessionId)
- [Update API Agent scheduled task](/docs/api-openapi/api-claw/patch-taskId)
- [Create Claw Agent](/docs/api-openapi/api-claw/post-agents)
- [Create API Agent](/docs/api-openapi/api-claw/post-api-agents)
- [Request API Agent session attachment upload](/docs/api-openapi/api-claw/post-attachments)
- [Connect API Agent channel](/docs/api-openapi/api-claw/post-channels)
- [Send a message to a Claw Agent](/docs/api-openapi/api-claw/post-chat-sessions)
- [Stream chat with an API Agent session](/docs/api-openapi/api-claw/post-chat): Starts or continues an API Agent chat using the same streaming logic as Buda's AI chat route. The response is an AI SDK UI message stream.
- [Confirm API Agent Drive upload](/docs/api-openapi/api-claw/post-confirm-upload)
- [Start managed channel connection](/docs/api-openapi/api-claw/post-connections)
- [Delete API Agent Drive item](/docs/api-openapi/api-claw/post-delete)
- [Disable API Agent scheduled task](/docs/api-openapi/api-claw/post-disable)
- [Get API Agent Drive download URL](/docs/api-openapi/api-claw/post-download-url)
- [Enable API Agent scheduled task](/docs/api-openapi/api-claw/post-enable)
- [Send API Agent session message](/docs/api-openapi/api-claw/post-messages)
- [Poll managed channel connection](/docs/api-openapi/api-claw/post-poll)
- [Rename API Agent Drive item](/docs/api-openapi/api-claw/post-rename)
- [Run API Agent scheduled task now](/docs/api-openapi/api-claw/post-run)
- [Create API Agent scheduled task](/docs/api-openapi/api-claw/post-scheduled-tasks)
- [Create API Agent session](/docs/api-openapi/api-claw/post-sessions)
- [Create Space](/docs/api-openapi/api-claw/post-spaces)
- [Read API Agent Drive text file](/docs/api-openapi/api-claw/post-text)
- [Get API Agent Drive upload URL](/docs/api-openapi/api-claw/post-upload-url)
