Tracecat is shipping like a team that expects AI to touch real security work, not just demo data. The website still describes it as an open source security automation platform for teams and AI agents, with over 200 pre-built integrations, MCP connectivity, scripts, loops, and human approvals. The recent releases match that positioning instead of contradicting it.
On May 1 alone, Tracecat cut beta.43, beta.44, and beta.45. That is not cosmetic churn. The notes cover action output masking, bearer auth for service account keys, service-account entitlement gating, workflow rollback, legacy preset compatibility, and a cutover in LLM providers. This is the boring infrastructure you need before AI-assisted security automation can survive production.
The product shape is clearer than the marketing copy. Tracecat says it supports remote and stdio MCP servers, parallel subflows, Python, Bash, and JavaScript, plus explicit approvals for human-in-the-loop steps. That points to controlled agent execution, not a chatbot pasted onto a workflow engine.
The 97 health score and 100 shipping score line up with that pace. Tracecat also says it is deployed at enterprise scale, which fits the release notes better than a consumer-style pitch.
What the data does not tell you
ToolVitals sees releases, stars, and cadence. It does not see code quality, customer satisfaction, revenue, or whether the product actually holds up during a real incident. The site makes strong claims, but this dataset does not prove adoption, operational reliability, or security outcomes.
So the story here is a plausible product direction, not a verdict.
Competitor context
Tracecat has 3,564 GitHub stars. Infisical has 26,416 stars and 21 release events in 30 days. Stack Auth has 6,770 stars and 27 release events in 30 days. Tracecat’s 11 release events in 30 days is still active, just not noisy.
That makes the signal simple. Tracecat is smaller than those peers, but it is clearly shipping, and the release notes show a team spending time on the hard parts, auth, masking, approvals, and model routing.
Bottom line
If your team wants open source security automation and plans to put AI inside the workflow, Tracecat deserves a real evaluation. If you only want a plain SOAR replacement, the agent-first framing is extra surface area you will need to justify.