Control coding-agent risk across your organization

Watcher monitors billions of agent tokens a month across production engineering teams at agent-building scale-ups, quant trading firms and multinational enterprises.

  • Prevent AI breaches that
    cost $6 million on average †
  • Catch malicious AI
    agents in real time
  • Secure without slowing
    down engineers
Book a call

Our research is trusted by

OpenAIAnthropic
Google DeepMind
Meta
Microsoft
The dark Watcher Analyzer showing two blocked attempts to change production Terraform outside the required review process.

Explore the full Watcher demo

How Watcher works

One set of policies. Visibility across your team’s coding agents.

Watcher deploys policies across your organization, checks agent actions before they run, and reviews sessions for incidents. Claude Code and Codex are supported; the open slot represents other integrations.

Secure policy rollout

100% of devices protected Deployed policy through MDM

Define your rules centrally and apply them across your engineering team.

Real-time blocking

blocked by Watcher terraform apply -auto-approve

Check actions before they run and block those that cross your risk threshold.

Your team’s coding agents

Claude Code
Codex
Others

Automated analysis

Recommended Action Update rule to prevent production bypass

Review sessions, investigate findings, and identify what needs attention.

Learn more

Catch serious failures.Keep agents moving.

Recall on critical-severity failures
100%
False positives on benign tool calls
<1%
Estimated added cost
3–5%
Estimated added latency
6–8%

Results for Watcher Live’s Default configuration. Detection and false positives measured on a synthetic attack dataset; cost and latency estimated from real agent traffic.

Read our latest announcement

Announcing Watcher Live

Frontier-grade agent security

Apollo Research is a leader in understanding AI scheming risks.
We apply that research to Watcher, so your team benefits from the expertise behind frontier model evaluations.

Apollo is the best place in the world for conducting scheming research. Extremely dedicated, hard working, collaborative. I’m glad OpenAI is working with them.
Wojciech Zaremba Co-founder, OpenAI
The joint OpenAI–Apollo investigations into model scheming are among the most consequential lines of research underway today. Apollo is an excellent place to do AGI safety work.
Jakub Pachocki Chief Scientist, OpenAI
Understanding deception and scheming in frontier models is essential for AI safety. Apollo Research’s work delivers insights the field urgently needs.
Yoshua Bengio Turing Award Laureate, most cited scientist in the world
Read Apollo’s research

Install Watcher

Watcher can be installed self-serve today or self-hosted.

curl -fsSL https://github.com/ApolloResearch/watcher-bin/releases/latest/download/install.sh | bash -s -- --cloud
# Restart or open a new terminal
watcher
Setup documentation

FAQ

What coding agents work with Watcher?
Watcher integrates directly with Claude Code and Codex. See the setup documentation to connect your agents.
Is Watcher cloud-hosted or self-hosted?
The install command on this page connects to Apollo’s managed monitoring platform in Western Europe. Organizations can also self-host Watcher in their own infrastructure; book a call to discuss deployment.
How is this different from auto-mode?
Watcher gives security teams centrally managed policies and visibility across developers using Claude Code and Codex, with monitoring models configurable independently of the agent. Read about organization-wide controls.

See how Watcher fits your team

Talk through your agents, security policies, and deployment needs.

Book a call