# Matimo Workbench

Workbench is a multi-tenant production platform for AI agents. You build an
agent, give it tools and knowledge, choose how it reasons, and deploy it where
your users are. It is not a developer framework. Cost tracking, audit logging
and tenant isolation come with it.

## What it does

- **Agents without code.** A visual agent builder with role templates,
  personality settings, and both short-term and long-term memory.
- **16 reasoning strategies.** Agents can reason with ReAct, Chain-of-Thought,
  Plan and Execute, Reflexion and others, instead of following a fixed
  flowchart. ReAct is available on every plan; all 16 are on Pro and Enterprise.
- **Knowledge.** Knowledge bases with retrieval over your documents, a knowledge
  graph, and connectors to Microsoft 365, Confluence and Notion.
- **Tools.** 129 ready-to-use tools through Matimo OSS, plus a governed catalog
  of 449 more. See [Integrations](/docs/integrations).
- **Your models, your keys.** Bring your own provider key. Matimo does not mark
  up model cost. Agents can fall back to a second provider if the first fails.
- **Cost control.** Cost tracking per execution, and optional per-agent budget
  caps that stop a run when a cost or token limit is reached.
- **Human approval.** An agent can pause before a consequential action and wait
  for a person to approve it.
- **Deployment.** Chat, the visual workflow editor ([Studio](/docs/studio)), a
  REST endpoint per agent, webhooks, and scheduled runs.

## What it does not do

- It does not run with zero human checkpoints by design. Approval gates are a
  core feature, not something to remove.
- It does not supply language models.

## Who it is for

Teams that want agents in production, not in a pilot, and need a platform their
security team can sign off on.

## Next step

Sign up free at [matimo.ai](/signup), or see the [API reference](/api) if you
plan to integrate programmatically.