5 fictional matters
Fictional litigation
Explore complaints, motions, discovery and oral argument across state and federal trial and appellate settings.
Put your argument
to the test.
An open-source legal AI workflow for drafting, challenging and refining legal work. Six simulated personas test the work from different sides, with a human attorney on the loop.
20 prepared examples. No account or API key needed to explore.
- Draft
Make the strongest case
for your position. - Challenge
What would opposing counsel
attack first? - Revise
Address the weakness.
Show what changed.
Assess, refine, review again.
Professional judgment stays with the attorney.
Start with a question that interests you.
Follow the draft, the challenge and the revision. See the reasoning and the unresolved issues along the way.
5 public-record cases
What if counsel took a different path?
Compare alternative strategies in high-profile cases, including Musk v. Altman and OpenAI. Examine how a different approach might have mattered.
10 business questions
Ask in-house counsel
Work through the questions business teams bring to their lawyers: supplier agreements, employment, data use and other everyday decisions.
These are prepared educational scripts, not live model runs or attorney-approved advice. Real-case alternatives are hypothetical, not predictions of what a court would have done.
More than a first draft.
Six personas take different roles in the review. You can inspect the work as it moves through the loop.
Draft and review
The Associate develops the work. The Partner reviews its support, framing and completeness.
Test the other side
Opposing-Counsel Associate and Partner challenge the position and expose weaknesses.
Revise and assess
The drafting team responds. The Judge and Rubric Judge assess the work against the task's criteria.
A human attorney on the loop.
Professional judgment has an explicit place in the workflow. Privilege decisions, admission dispositions and attorney attestation remain human responsibilities.
See the work behind the answer.
Trace the revisions
Read drafts, opposing critiques and revised work together. Examine how the response to a challenge changes the result.
Inspect the gates
See rubric results, unresolved support and citation issues, and decisions that still need attorney review.
Replay an example
Download a demo's inputs and replay its prepared workflow locally, without live model calls.
Build on open source
Inspect task adapters, persona workflows, budget controls and audit trails. Extend the project for your own experiments.
For people who work with the law.
And people who want to understand it.
- Lawyers
- Explore how a position holds up under opposing review.
- Law students
- Compare arguments and study the choices behind a revision.
- In-house counsel & legal operations
- Follow business questions through a structured review process.
- Developers
- Study an open-source legal multi-agent workflow and its controls.
- Anyone curious about the law
- See how a legal question can lead to competing arguments.
Explore here.
Make it your own locally.
The public site is a read-only viewer. To replay examples or work with your own materials, install MootLoop on your computer.
Start with Python 3.12 and uv. The local-use guide walks through installation, demo downloads and replay commands.
Read the local-use guideGet the source
git clone https://github.com/damienriehl/mootloop.git
cd mootloop
uv sync
uv run mootloop --helpKeep matter files outside the repository and background-sync folders. Prepared replays require no provider key; other workflows have their own setup requirements.