The verifier in your agent's loop
Everything for wiring Verfix into an AI coding agent's edit → verify → fix loop — CLI reference, config schema, execution modes, and the config-first guardrails that keep it from rewriting your app.
New to Verfix? Let's install the CLI
Install the CLI, run `verfix init`, and run your first browser verification flow in under 2 minutes — no Docker required.
API & References
CLI Reference
Command reference manual for managing servers, initiating runs, and reviewing execution flows via standard shell scripts.
Read CLI guide →Config Reference
Complete configuration parameters for `verfix.config.json` including timeout settings, flow parameters, and system schemas.
See config specs →Concepts & Architecture
Execution Modes
Learn the behaviors, guarantees, and trade-offs of the Strict, Assisted, and Exploratory execution strategies.
Understand run styles →Failure Taxonomy
Explore the full categorization of failure types, including visibility issues, selector errors, and structured hint layouts.
View taxomony schema →Developer Guides
Agent Integration
Step-by-step guides for connecting Verfix structured JSON feedback into LLM architectures, AutoGen, and LangChain environments.
Connect AI Agents →Config-First Verification
Why agents should fix flows, not rewrite your app, and the git-baseline source guard that enforces it.
Read the guide →Self-Hosting Guide (Optional)
The opt-in Docker server runtime — API, queue, and timeline dashboard — for teams that want a shared deployment.
Deploy to private servers →