# Runbear Docs

> Current documentation for building, deploying, and operating Runbear AI agents.

Use this index to select the smallest relevant Markdown page. Guides for the retired OpenAI Assistant and GPT import paths are intentionally omitted; /legacy.md covers what replaced them.

## Start here

- [Runbear overview](https://docs.runbear.io/index.md): See what an agent is made of and where it can be deployed.
- [Getting started](https://docs.runbear.io/getting-started.md): Create an agent and deploy it to Slack.
- [Inbox Agent](https://docs.runbear.io/inbox-agent/overview.md): Set up the personal assistant that drafts replies for review.

## Build an agent

- [Instructions](https://docs.runbear.io/agents/instructions.md): Define agent behavior in named context blocks.
- [Knowledge](https://docs.runbear.io/agents/knowledge/overview.md): Ground answers in documents synced from your tools.
- [Notion knowledge](https://docs.runbear.io/agents/knowledge/notion.md): Sync Notion pages and databases.
- [Confluence knowledge](https://docs.runbear.io/agents/knowledge/confluence.md): Sync Confluence spaces and pages.
- [Google Drive knowledge](https://docs.runbear.io/agents/knowledge/google-drive.md): Sync Drive files by OAuth or a service account.
- [Slack knowledge](https://docs.runbear.io/agents/knowledge/slack.md): Sync messages from selected Slack channels.
- [SharePoint knowledge](https://docs.runbear.io/agents/knowledge/sharepoint.md): Sync SharePoint documents and pages.
- [Website knowledge](https://docs.runbear.io/agents/knowledge/website.md): Crawl a public website into an agent's knowledge.
- [Local files](https://docs.runbear.io/agents/knowledge/local-files.md): Upload files directly into an agent's knowledge.
- [Tools](https://docs.runbear.io/agents/tools/overview.md): Let an agent act and read live data over MCP.
- [Tool catalog](https://docs.runbear.io/agents/tools/catalog.md): Browse the managed MCP servers Runbear hosts.
- [Custom MCP](https://docs.runbear.io/agents/tools/custom-mcp.md): Connect your own MCP server to an agent.
- [Zoom](https://docs.runbear.io/agents/tools/zoom.md): Connect Zoom's official MCP server.
- [Memory](https://docs.runbear.io/agents/memory.md): Configure what an agent remembers between conversations.
- [Testing and improving](https://docs.runbear.io/agents/testing.md): Try an agent privately and fix what it gets wrong.
- [Agent settings](https://docs.runbear.io/agents/settings.md): Choose the model, reasoning effort, access, and data handling.

## Deploy and automate

- [Slack](https://docs.runbear.io/channels/slack.md): Answer mentions and direct messages in Slack.
- [Microsoft Teams](https://docs.runbear.io/channels/microsoft-teams.md): Deploy to Teams, optionally under your own branding.
- [Discord](https://docs.runbear.io/channels/discord.md): Deploy an agent to a Discord server.
- [HubSpot](https://docs.runbear.io/channels/hubspot.md): Assist with customer conversations in the HubSpot inbox.
- [Zendesk](https://docs.runbear.io/channels/zendesk.md): Draft internal notes or public replies on tickets.
- [Triggers](https://docs.runbear.io/triggers.md): Run an agent on a schedule, a webhook, or an app event.

## Inbox Agent

- [Inbox Slack](https://docs.runbear.io/inbox-agent/channels/slack.md): Get suggested replies when you are mentioned.
- [Inbox email](https://docs.runbear.io/inbox-agent/channels/email.md): Get draft replies for incoming Gmail messages.
- [Inbox calendar](https://docs.runbear.io/inbox-agent/channels/calendar.md): Get a prepared brief before each meeting.
- [Inbox tools](https://docs.runbear.io/inbox-agent/tools.md): Give Inbox Agent internal context to draft from.

## Developer reference

- [API keys](https://docs.runbear.io/api/api-keys.md): Create and revoke keys for the REST API.
- [Web SDK](https://docs.runbear.io/api/web-sdk.md): Embed a Runbear agent in a React application.
- [MCP server](https://docs.runbear.io/api/mcp-server.md): Manage Runbear agents from an OAuth-capable MCP client.
- [Agent Skills](https://docs.runbear.io/api/agent-skills.md): Deploy local agent projects from Claude Code.
- [Traces API](https://docs.runbear.io/api/traces.md): Inspect agent execution traces programmatically.
- [OpenAPI specification](https://api.runbear.io/v1/openapi.json): Machine-readable schema for the public REST API.

## Optional

- [Members and roles](https://docs.runbear.io/administration/members-and-roles.md): Control who can see and change what.
- [Tool approvals](https://docs.runbear.io/administration/tool-approvals.md): Require review before an agent may use a tool.
- [Single sign-on](https://docs.runbear.io/administration/sso.md): Authenticate members through your identity provider.
- [Audit log](https://docs.runbear.io/administration/audit-log.md): Review configuration changes in the organization.
- [Billing and credits](https://docs.runbear.io/administration/billing-and-credits.md): Understand credit consumption and interactors.
- [Analytics](https://docs.runbear.io/administration/analytics.md): Review agent adoption and feedback.
- [Legacy agent types](https://docs.runbear.io/legacy.md): See what changed for retired OpenAI Assistant agents.
- [FAQs](https://docs.runbear.io/faq.md): Read security, data-handling, and pricing answers.