Skip to content
@maida-ai

Maida AI

Maida.AI

Don't let broken agent changes merge.

Maida is the pre-merge behavioral regression gate for AI agents. It records agent runs, compares current behavior against a known-good baseline, and fails CI when structural behavior regresses: more steps, unexpected tool calls, loops, latency spikes, or cost blowups.

Maida sits earlier than production tools and complements output evals. Output evals ask whether the answer was good. Maida asks whether this PR changed how the agent behaves.

No cloud account required. No telemetry by default. Runs stay in your environment unless you explicitly export them.


Repositories

Repo What it is
maida Python package, maida CLI, SDK, local run storage, and timeline viewer
maida-assert GitHub Action for running Maida behavioral checks on pull requests
maida-tutorials Example workflows and walkthroughs for trying Maida on agent code
maida-ts TypeScript compatibility layer for integrations and plugin authors
opencode-plugin Maida plugin for OpenCode

Quick Start

Install the maida-ai package from PyPI. It provides the maida CLI and the maida Python module.

pip install maida-ai

Instrument one agent entrypoint:

from maida import trace

@trace
def run_agent(input: str):
    # Your existing agent code
    ...

Capture a known-good baseline:

python my_agent.py
maida list --json
maida baseline <RUN_ID> --out baselines/my_agent.json

Assert a future run against that baseline:

python my_agent.py
maida assert <RUN_ID> --baseline baselines/my_agent.json --policy .maida/policy.yaml --format markdown

GitHub Action

Use maida-ai/maida-assert@v2 to block behavioral regressions before merge.

name: Agent Regression Check

on: [pull_request]

jobs:
  agent-check:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
    steps:
      - uses: actions/checkout@v4
      - uses: maida-ai/maida-assert@v2
        with:
          agent-script: my_agent.py
          baseline: baselines/my_agent.json
          policy: .maida/policy.yaml
          python-version: '3.12'

Your agent-script must use @trace or traced_run() so Maida can record a run.


What Maida Flags

  • Step-count regressions
  • Tool-call count regressions
  • Unexpected tool paths
  • Loop risk
  • Guardrail events
  • Latency envelope changes
  • Cost envelope changes
  • Missing stop conditions

These are behavioral regression signals. They do not prove output quality. They tell you when an agent's execution behavior changed relative to a baseline.


Website: maida.ai PyPI: maida-ai Contact: contact@maida.ai

Popular repositories Loading

  1. maida maida Public

    Your agent still returns the right answer -- but now it calls 3x the tools. Maida is the pre-merge behavioral regression gate for AI agents: baseline agent traces, gate PRs in CI, block behavior re…

    Python 31 7

  2. pymai pymai Public archive

    Tiny‑yet‑powerful Python framework for composable AI agents and durable workflows.

    Python 5

  3. maida-ts maida-ts Public

    TypeScript mirror of the Maida trace contract. Emit and validate Maida-native agent traces from Node. Deliberately narrow -- not at feature parity with the Python package.

    TypeScript 5

  4. maida-assert maida-assert Public

    GitHub Action for the Maida pre-merge behavioral regression gate. Runs your traced agent in isolated trials on every PR, compares the runs against a baseline and policy, and posts a sticky Markdown…

    Python 4

  5. XCP XCP Public archive

    eXtensible Coordination Protocol

    Python 2

  6. MaiEther MaiEther Public archive

    Communication medium for cross-agent comms

    Python 2

Repositories

Showing 10 of 16 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…