> ## Documentation Index
> Fetch the complete documentation index at: https://comis-fix-skill-import-vetting-gate.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Welcome to Comis

> Open-source security-first runtime for AI agents that learn and act across sessions

> **Let agents learn and act. Keep authority in the runtime.**

**Governed learning. Bounded action. Recorded evidence.**

Comis is an Apache-2.0, self-hosted, security-first runtime for AI agents that
learn and act across sessions. It is built for agents that work on a schedule,
use real tools, build up context, or need to recover after a failure.

The promise is simple: learned guidance can influence what an agent proposes,
but it cannot grant permission. Comis keeps identity, capabilities, credential
scope, tool policy, and configured limits in the runtime, outside prompt text.
It also keeps important run state and bounded evidence available for recovery
and review.

Comis is under active development. Start with one agent and one limited
workflow. Test the controls on your host before you add broader access.

## Why Comis

Comis is designed for work that continues beyond one chat window.

<CardGroup cols={3}>
  <Card title="Keep authority outside the model" icon="shield-check" href="/get-started/security">
    Apply capability, origin, credential, tool, and configured budget controls
    in runtime paths rather than relying on prompt instructions.
  </Card>

  <Card title="Govern learning across sessions" icon="brain" href="/agents/memory">
    Use source records, trust signals, configured corroboration, usefulness, and
    correction history to govern which experience can return as guidance.
  </Card>

  <Card title="Recover and explain the work" icon="magnifying-glass" href="/operations/observability">
    Recover selected context and review run outcome, cost, policy failures, and
    a rule-based likely cause when the recorded evidence matches one.
  </Card>
</CardGroup>

## What the runtime brings together

<CardGroup cols={3}>
  <Card title="Long-running workflows" icon="robot" href="/agents/execution-graphs">
    Coordinate sequential and parallel work with dependencies, retries,
    budgets, and configured node-boundary recovery.
  </Card>

  <Card title="Learned memory with history" icon="puzzle-piece" href="/agents/memory">
    Record sources and trust signals, require configured evidence, and preserve
    prior history when eligible learned state is corrected. Learned text can
    guide proposals, but it cannot expand runtime authority.
  </Card>

  <Card title="Channels and interfaces" icon="messages" href="/channels/index">
    Start work from schedules, messaging channels, the web dashboard, CLI, or
    APIs, then inspect it through the same runtime.
  </Card>
</CardGroup>

## Start here

```bash theme={}
npm install -g comisai
comis init
comis daemon start
```

The setup wizard defaults to encrypted credential storage. Back up the
generated master key separately, then follow the [Quickstart](/get-started/quickstart)
to verify the daemon, your first agent, and the security posture.

Comis runs on infrastructure you control. Configured model, embedding, channel,
media, and tool providers can still receive data routed to them.

## Explore the documentation

<CardGroup cols={2}>
  <Card title="How It Works" icon="diagram-project" href="/get-started/how-it-works">
    Follow a request through routing, execution, controls, state, and evidence.
  </Card>

  <Card title="Use Cases" icon="lightbulb" href="/get-started/use-cases">
    Start with an overnight research brief or a limited, read-only investigation.
  </Card>

  <Card title="Security Boundaries" icon="shield-check" href="/get-started/security">
    Review defaults, opt-in controls, deployment assumptions, and residual risk.
  </Card>

  <Card title="Operations" icon="gauge" href="/operations/index">
    Deploy, monitor, back up, recover, and troubleshoot the daemon.
  </Card>
</CardGroup>