Skip to main content

​
What is Turen?

Turen gives your organization visibility and control over AI coding agents like Claude Code, without slowing developers down. It runs a lightweight agent on each developer’s machine (macOS, Linux, or Windows) to enforce security rules, record sessions, and report telemetry.
Turen dashboard showing a live Claude Code terminal flagging a chalk typosquat on the left, and a Claude Code Activity panel on the right with 247 packages scanned, 12 blocked, and 3 SAST findings.

One pane of glass over Claude Code: the live session on the left, Turen's security findings (packages scanned, blocks, SAST findings) on the right.

Built-In Rules

Block dangerous commands, prompt injections, and secret exposure before they happen.

Session Replay

Record and replay every Claude Code session with full analytics and searchability.

LLM Telemetry

Track token usage, costs, latency, and model usage across your entire team.

Fleet Management

Monitor every agent machine, enforce policies centrally, and manage access at scale.

​
Who is Turen for?

Security teams who need guardrails around AI agent usage: blocking dangerous commands, detecting prompt injection attempts, and preventing secret leakage. Engineering leaders who want visibility into how AI agents are being used: what code is being generated, what tools are being invoked, and how much it costs. Compliance teams need an audit trail of every AI interaction for regulatory requirements.

​
How it works

Turen deploys a lightweight agent on each developer’s machine. Every installer also bundles the Turen desktop app. The agent includes:
  1. A background daemon that registers with your Turen organization, syncs policies, uploads session data, and derives usage telemetry from Claude Code’s session transcripts.
  2. Claude Code hooks that enforce your security rules locally: validating commands, gating package installs, and scanning code as it’s written.
To set a machine up, install the agent and sign in through the Turen desktop app. That registers the machine, so no registration key is required for developer installs. Keys remain available for headless and fleet rollouts. Claude Code connects directly to the Anthropic API, and developers use it exactly as before. Everything else is automatic.

​
Platform Support

Codex support is in preview. Session replay, LLM telemetry, package security, and skill delivery behave the same as they do for Claude Code, and plan-limit reporting and command, injection, and secret activity now surface for Codex too. Local rule enforcement runs through Codex hooks but does not yet cover every tool path, so treat it as an added layer of defense rather than a complete boundary.

Ready to get started?

Set up Turen in under 5 minutes.