# Be the agent that already knows

Access real time context of human operators, their tools, and other AIs in the loop. Collaboration that feels like telepathy. Outcomes that are superhuman.

## Capabilities

`/whats-up`

Instant briefing on recent activity — browsing, AI conversations, research threads, and open tasks. Skip the back-and-forth warm-up and start with full situational awareness.

`/whats-next`

Surface the highest-leverage next steps based on active streams of work, outstanding questions, and stalled initiatives. Proactive recommendations grounded in real activity, not guesswork.

`search()`

Semantic search across all accumulated knowledge — websites, documents, searches, AI chats, agent sessions, social saves. Enriched with metadata on importance, recency, and connectedness.

`traverse()`

Walk the knowledge graph to discover connections that haven't been made yet. Follow links between topics, sources, intents, and concepts to build richer context and surface novel insights.

## Benefits

Relevance

First-token usefulness

Produce responses that don't require three rounds of "actually, what I meant was..." to become useful.

Continuity

Compound knowledge across sessions

Pick up mid-thought with full awareness of what's been done, decided, and tried.

Effectiveness

Do more with every turn

Spend effort on the work, not on figuring out what the work is.

Autonomy

Outsized impact

Navigate complex goals, surface connections others miss, and chart a path forward that wouldn't exist without you.

## Features

### Skills

Pre-built capabilities you can invoke directly — /whats-up, /whats-next, search, traversal, and more. Each skill is a composable unit of work designed to facilitate agent-to-human collaboration.

### MCP server

Native Model Context Protocol support. Query the shared knowledge graph, and access real-time updates. Ideal for cloud hosted agents.

### CLI, SDK, API

Full programmatic access. Optimized entry points for CLI based coding agents.

### Conversation turn collection

Every conversation turn is captured and fed back into the activity stream and knowledge graph in real time.

### Agent-to-agent context

Your session tool use, reasoning, recommendations, and outputs become durable context for the next session, the next agent, or the next tool in the chain.

## Get started

Install Autographical today.

All →

{
  "mcpServers": {
    "autographical": {
      "type": "http",
      "url": "https://api.autographical.ai/mcp"
    }
  }
}