# wirelatch

> wirelatch is an agent gateway, sandbox and audit trail in one. It connects AI agents to a company's code, tools and data without sharing keys, runs each agent in its own isolated sandbox, and keeps an append-only record of every action, ready for audits.

wirelatch is in early access and is onboarding a small group of design partners.

Teams use wirelatch to:

- Give AI agents (Claude Code, Codex, Gemini CLI, MCP-based or custom agents) access to company systems without handing over passwords, tokens or API keys.
- Deploy agents into isolated sandbox runtimes instead of laptops or shared servers.
- Put a gateway, including an MCP gateway, between agents and tools: allow what each agent may reach, block the rest, and send sensitive actions for review.
- See what every agent is doing, live, and keep an audit trail that security teams, auditors and regulators can review.

## Product

- [Agent gateway](https://wirelatch.dev/#product): Connect agents to GitHub, Slack, Linear, Jira, Postgres, MCP servers and your own APIs through one gateway. Credentials stay in wirelatch; each agent gets only the access you allow.
- [Agent sandbox](https://wirelatch.dev/#product): Every agent runs in a sealed sandbox of its own. Pause it, bring it back, and it picks up where it left off.
- [Audit trail](https://wirelatch.dev/#audit): Every action is recorded as it happens: what the agent did, which system it touched, and whether it was allowed, blocked or waiting for review. Export the record for auditors.

## Docs

- [wirelatch in plain text](https://wirelatch.dev/index.html.md): The problem, the product, audits and how it works, plus answers to common questions: what an agent gateway is, how to sandbox and deploy AI agents, how to give them access without sharing keys, and how to audit what they did.

## Optional

- [Join early access](https://wirelatch.dev/#join): Leave a work email to join the design partner list.
- [Full text bundle](https://wirelatch.dev/llms-full.txt): Same content as /index.html.md.