# HypAware

> HypAware is the improvement loop for all your agents. It reads every agent session and shows what's broken, the evidence, the fix, and whether it stayed fixed. Agents can query their own past sessions with `hyp query` and act on a recommendation by its ID. Install with `npm i -g hypaware`, then `hyp setup`. Records Claude Code, Claude Desktop, Codex, Cursor, OpenCode, Pi, Factory Droid (local CLI transcripts), OpenClaw, Hermes Agent, and Ollama, with VS Code (GitHub Copilot), LangGraph (JavaScript and Python callbacks), and Codex Cloud in beta. Requires Node.js 22.12 or newer on macOS or Linux. Open source (MIT): https://github.com/hyparam/hypaware

## Start
- [Quickstart](https://hypaware.ai/quickstart/index.md): install, confirm capture, query your history, connect to your team, privacy controls
- [Home](https://hypaware.ai/index.md): what HypAware does and how the loop works
- [Integrations](https://hypaware.ai/integrations/index.md): every agent and tool HypAware works with, grouped by kind (coding agents, cloud agents, agent harnesses, software factories and CI, local models, developer tools), with status
- [Agent observability](https://hypaware.ai/observability/index.md): See every session from Claude Code, Codex, and Cursor. Then HypAware finds what's broken, hands back the fix, and checks that it stayed fixed.

## Docs
- [Documentation index](https://hypaware.ai/docs/index.md)
- [Setup](https://hypaware.ai/docs/cli/index.md): installation, updates, background service, uninstall
- [Integrations](https://hypaware.ai/docs/integrations/index.md): every supported agent and tool, its status (supported or beta), and setup for each: CI and headless hosts, VS Code GitHub Copilot (local single-folder workspaces) and LangGraph (JavaScript and Python callbacks), plus Codex Cloud, Ollama, and GitHub
- [Clients and history](https://hypaware.ai/docs/clients/index.md): choosing clients, capture, history imports, detaching, skills
- [Queries](https://hypaware.ai/docs/querying/index.md): ask, search, SQL, graph and vector tools, MCP
- [MCP](https://hypaware.ai/docs/mcp/index.md): connect an MCP client, locally with hyp mcp serve or remotely with an agent token
- [MCP reference](https://hypaware.ai/docs/mcp-reference/index.md): MCP tools and prompts, with their CLI equivalents
- [Reports](https://hypaware.ai/docs/reports/index.md): generate, review, publish
- [Cloud and teams](https://hypaware.ai/docs/team-setup/index.md): enrollment, sync, remote access
- [Privacy](https://hypaware.ai/docs/privacy/index.md): what is recorded, sharing controls, exclusions, deletion, telemetry
- [Configuration and storage](https://hypaware.ai/docs/configuration/index.md): config files, retention, cache, exports, plugins
- [Troubleshooting](https://hypaware.ai/docs/troubleshooting/index.md): missing recordings, sync failures, install problems
- [CLI reference](https://hypaware.ai/docs/cli-reference/index.md): every command, flag, and alias
- [Plugin authoring](https://hypaware.ai/docs/plugin-authoring/index.md): for contributors

## Optional
- [Full text of the site and docs](https://hypaware.ai/llms-full.txt)
- [Pricing](https://hypaware.ai/pricing)
- [Security](https://hypaware.ai/security)
- [Privacy Policy](https://hypaware.ai/privacy)