The behavior layer for AI agents

You review the diff. Nobody reviews the behavior.

Maida shows how your agent behaves, and fails the pull request check when that behavior changes in ways you didn't approve.

Baseline and pull request execution trajectories A known-good execution and a pull request execution begin at the same input and agent, then diverge. The pull request repeats search and introduces a new CRM tool before returning the same answer. KNOWN GOOD PR CHANGE input A agent search lookup answer search search search NEW CRM same answer
Baseline Changed in PR Same input. Same answer. Different behavior.

02 The incident

Same answer.
Different behavior.

Your agent still returns the right answer, but now it calls 3x the tools. Your output checks pass, even though the way your agent gets there has changed.

03 The behavior

How your agent actually behaves.

Which tools it calls and in what order. How many calls and steps it takes, where it retries, where it risks a loop. How latency and cost move with all of it. That is the behavior you review, and the behavior every later change is held to.

Illustrative comparison Baseline / candidate fingerprint
Baseline Candidate
Illustrative baseline and candidate agent behavior metrics
Signal Baseline Candidate Shape change
Tool sequence lookup -> answer lookup x4 -> CRM -> answer new path
Tool calls25 +150%
Steps48 2x
Retries02 appeared
Loop risknonedetected flagged
Latency1.2 s3.8 s +217%
Cost$0.006$0.019 +217%

Accept change on purpose. You decide which behavior changes are acceptable. Maida compares each run with the behavior you approved, and nothing else.

04 The verdict

The verdict on your pull request.

PASS
The observed behavior meets your approved checks.
FAIL
The observed behavior violates your approved checks.
INCONCLUSIVE
There is not enough evidence to decide; this is never an approval.
Maida Demo PR-comment excerpt

❌ Maida verdict: fail

5 of 7 checks failed

Top behavior changes

BehaviorBaselineCurrentChange
Tool calls37+133%
  • Repeated tool call

    search_kb — repeated 1 -> 5 calls

    tool_call_count_exceeded
  • New tool

    escalate_to_human — new tool, not in baseline

    new_tool_path
  • send_reply — no longer called

05 Runs where your code lives

Runs on your machine or CI runner. No Maida cloud account required.

Maida's evaluation runs locally and does not upload raw traces to Maida. Your agent's own model and tool calls are a separate flow Maida does not create or remove. Anything you choose to enable states where it sends data and what it sends.

Inspect it. Open the local execution timeline.

Archive it. Keep traces with your own CI artifacts.

Control it. Export only when you explicitly choose to.

06 Install

Install Maida.

  1. 01
    Capture one task

    Use your coding agent in your repository.

  2. 02
    Review a few checks

    Choose the observed behavior you want to protect.

  3. 03
    Try a change

    See a pass, a deliberate failure, and repair.

  4. 04
    Add CI when ready

    Verify the repository boundary on a real PR.

Try the complete story first

No clone. No API keys.

A deterministic demo goes from a healthy trace to a failed check.

$ uv tool install "maida-ai==0.6.1"
$ maida demo --regression
Open quickstart

Choose the behavior you want to protect.

A short policy example from the pinned release docs:

version: 2
trials: 3
fail_fast: true
metrics:
  stop_condition_reached: {kind: invariant, require: true}
  step_count:
    kind: measured
    direction: upper
    tolerance: {relative: 0.5}
Read the policy reference

Add the check to your pull request.

For a traced Python entrypoint with a reviewed baseline:

maida run my_agent.py --baseline .maida/baselines/my_agent.json \
  --policy .maida/policy.yaml --format markdown

Follow the Action setup to configure a required PR check that reads the verdict.

GitHub Action: maida-ai/maida-assert@v0.6.0

Check permission: checks: write

See the broken PR demo Docs

07 Where this goes

Delegate the work. Keep the responsibility.

Agents will do more work than people can watch. That only works if their behavior is understood well enough to be trusted in advance, not inspected after. Maida starts at the pull request, where a change can still be stopped. We are building toward a standard for how agents should behave, on any model and any harness.

  1. 01
    The pull request gate

    A change in agent behavior is caught where it can still be stopped.

  2. 02
    Accept change on purpose

    Every change in behavior is reviewed by a person and accepted or rejected, on the record.

  3. 03
    Understand how agents decide

    Maida will see enough behavior, closely enough, to predict where it breaks and how to use it well.

  4. 04
    A standard for agent behavior

    Maida will state what an agent should do, on any model and any harness, before it runs.

Protecting an agent in a setup we haven't shown? Tell us what it does. contact@maida.ai