# TARX Docs

TARX is an AI work system for controlled context, tools, memory, evidence, and runtime choice.

## Core concepts

- Spaces hold context for a body of work.
- Chat does work using Space context.
- Comms handles people, drafts, handoffs, and approvals.
- Computer is the user's local machine and tools.
- Control governs account, memory, Vault, apps, approvals, billing, and capacity.
- Evidence shows what TARX used, changed, or decided.
- Supercomputer is optional hosted capacity, not the default product path.

## Adoption touchpoints

- Getting Started: https://docs.tarx.com/users/overview
- Connections: https://docs.tarx.com/developers/connections
- MCP: https://docs.tarx.com/developers/mcp
- Claude Desktop MCP: https://docs.tarx.com/developers/mcp-claude-desktop
- OpenAI Agents MCP: https://docs.tarx.com/developers/mcp-openai-agents
- API: https://docs.tarx.com/developers/api
- Enterprise evaluation: https://docs.tarx.com/enterprise/evaluation
- Provider and model strategy: https://docs.tarx.com/enterprise/provider-and-model-strategy
- Usage and capacity: https://docs.tarx.com/pricing/overview

## Capability status labels

- Available now: the surface exists and can be used or reviewed today.
- Pilot-ready: suitable for controlled evaluation or early account access.
- Enterprise evaluation: requires enterprise review, agreement, or implementation planning.
- Government evaluation: requires public-sector review, approval path, or implementation planning.
- Roadmap: directional, not available for use.

## Contact

Email: howdy@tarx.com
Slack: https://join.slack.com/t/tarxxxxx/shared_invite/zt-2aphbjypj-IrWza6yDP3oL8yxfYy4gPA