About Agent Jig
Agent Jig is an evaluation and testing framework for AI agents. Teams define test scenarios in YAML, run them against their agent in CI, and get deterministic pass/fail scores and regression diffs on every deploy.
The product solves the "silent regression" problem: AI agents that degrade silently after model updates, prompt changes, or framework upgrades, without anyone noticing until users complain.
Company boilerplate
Agent Jig is a developer tool for evaluating and testing AI agents. The platform enables engineering teams to define test scenarios in YAML, run them in CI pipelines, and get deterministic pass/fail scores and regression diffs on every deploy. Agent Jig supports all major agent frameworks and CI systems. The Free plan includes 50 evals per month; Pro starts at $89/month.
Key facts
- Founded 2026
- Product: AI agent evaluation framework (YAML-based, CI-native)
- Pricing: Free (50 evals/mo), Pro ($89/mo, 5,000 evals + CI integration), Enterprise (custom)
- Frameworks supported: LangChain, LlamaIndex, CrewAI, AutoGen, custom Python/Node, and any HTTP endpoint
- Website: agentjig.com
Press contact
For press inquiries, interviews, or additional information, contact [email protected]. We respond within one business day.