Testim automates web, mobile, and Salesforce app testing with AI, for agile teams and Salesforce practitioners.
Free plan · Paid plans: see official pricing · Free trial
Tracely
Tracely traces AI agent runs, grades them with evaluators, and turns production failures into regression tests that gate pull requests.
- Paid
- API · Plugin or MCP
- by Julien Wuthrich
- Verified 20 Oct 2026
What Tracely does
Records AI agent runs and grades each trace as it arrives. Converts failing production runs into tests that block pull requests from merging.
Built for
- Teams running AI agents who gate releases on production failures
Key features
- Streams agent traces over OTLP into a dashboard
- Online evaluators grade traces as they arrive
- One click freezes a failing trace into a replayable case
- Replays suite on each pull request and exits non-zero on regression
- Groups failing traces into named issues
- Judge calibration against human review labels
- MCP server for coding agents to read traces
Pros
Pros
- The core product is available under an MIT license for self-hosting.
- Instrumentation needs only a few lines of code added to an agent.
Where Tracely is listed
Tracely is listed under AI testing tools in AI coding tools.
See MCP servers for coding and Claude skills for coding.
DebuggingDevOps and infrastructureData analysis
Tracely alternatives
Compare all 5
testRigor runs end-to-end UI tests written in plain English, for QA teams, manual testers and engineering organizations.
Free plan · Paid plans: see official pricing · 14-day free trial
Momentic tests web and mobile apps end to end using plain-English tests, for engineering teams shipping every day.
Free plan · Paid from $125/mo
mabl is an agentic testing platform that validates web, mobile, and API experiences, for software engineering and QA teams.
Contact sales · 14-day free trial
Applitools tests user interfaces with Visual AI and automates test upkeep, for developers, QA staff and enterprise teams.
Paid from $667/mo (billed annually)