§

Local legal-writing
practice engine for law students

Generate model IRAC, CRAC, CREAC, and IRREAC answers, grade your own writing, drill MBE questions, spot issues, and create fresh hypos — with instant AI feedback on every answer.

Get Started Free → See the modes
💻 Runs on any laptop 💸 Free forever 📚 12 subject areas ⚡ Instant feedback

Law students practice the slow way:

  • Write essays in a vacuum with no feedback
  • Pay $200+/mo for bar prep AI add-ons
  • Wait days for a practice essay to get marked
  • Generic AI conflates Rule Statement with Rule Exploration
How IRAC Coach fixes that

Five modes. One AI engine.

Pick an area of law — Contracts, Torts, ConLaw, Crim, Property, Evidence, and more. Then pick what you need right now.

Generate

— the AI writes the analysis so you can study it

Paste any fact pattern or case summary. Get a structured IRREAC response with citations, element breakdowns, and study tips — in about 30 seconds.

  • IRAC AnalysisFull Issue → Rule Statement → Rule Exploration → Application → Conclusion, with real citations. Streaming output so you see it build.
  • Hypo GeneratorAuto-write fresh MEE-style fact patterns scoped to any topic. Never run out of practice material.

Practice

— you do the work, the AI grades every line

Every mode can generate its own fact pattern, so you don't need to bring your own. Pick area, pick topic, and the LLM writes a fresh hypo instantly.

  • Issue SpottingList every issue you see. AI scores coverage — what you caught, what you missed, and any false alarms.
  • MBE PracticeOne MBE-style question on demand. Every answer choice explained, not just the correct one.
  • Practice & FeedbackWrite section-by-section IRAC or a full essay. Graded against the AI's model answer with specific gaps highlighted.

Outlines

— rule references for any subject

Upload your own outlines (TXT/Markdown) or let the AI generate a basic one on first use. Cached for instant reuse.

  • Bring Your OwnDrop Emanuel, Glannon, or your professor's notes into default_outlines/. The app reads them verbatim.
  • Auto-Generated FallbackNo outline file? The LLM builds a rule reference on first request and caches it for next time.
The secret weapon

IRREAC — the split that earns A's

Standard IRAC compresses the whole rule into one step. IRREAC splits it into Rule Statement and Rule Exploration — and that split is worth exam points.

I

I — Issue

Frame the precise legal question the court must answer.
Whether [X] given [key facts].
R1

R1 — Rule Statement

State the applicable rule with a specific citation — Restatement §, UCC §, or landmark case.
Citation is mandatory. "The law requires…" earns no credit.
R2

R2 — Rule Exploration

Grade booster
How have courts interpreted the rule? Key cases, majority vs minority, circuit splits.
This is what separates a B answer from an A.
A

A — Application

Break the rule into elements. Analyze each separately against the specific facts. Include counter-arguments.
~50% of your exam score lives here.
C

C — Conclusion

Answer the Issue directly. Add a confidence level (High / Moderate / Low) with a one-sentence reason.
Never introduce new analysis here.
💡 The two-Rule split (R1 + R2) is the key upgrade. Rule Exploration — how courts have interpreted the rule, key cases, majority vs minority views — is what separates a B answer from an A.
Get running

Clone, run, done.

1

Clone and run

git clone https://codeberg.org/russkysong/IRAC-coach.git
cd IRAC-coach
bash setup.sh
source .venv/bin/activate
python launch.py

Then open http://localhost:8787. setup.sh installs the app and its dependencies. Later launches skip setup and start straight away.

For agent users

MCP connector (optional)

Connect IRAC Coach to Claude Code, Cursor, or any Codex-style agent. 

python mcp_server.py

Seven tools the agent can call:

  • irac_generateStructured IRAC/CRAC/CREAC/IRREAC from a fact pattern.
  • irac_gradeGrade a section-by-section answer against a model answer.
  • irac_grade_essayGrade a full essay against a model answer.
  • irac_issue_spotScore issue coverage — caught, missed, and extra.
  • irac_hypoGenerate a fresh MEE-style fact pattern on demand.
  • irac_mbeGenerate an original MBE question with per-choice explanations.
  • irac_outline_summarySummarise a rule outline for a topic.

⚠️ Privacy note

In MCP mode the host agent can see any facts, answers, and generated responses you send to a tool. Commercial outlines are not injected by default.