brew install argon-lab/tap/argonctl      # macOS
npm install -g argonctl                  # cross-platform

# MongoDB must run as a replica set (one-node is fine):
docker run -d --name argon-mongo -p 127.0.0.1:27017:27017 mongo:7 --replSet rs0
docker exec argon-mongo mongosh --quiet --eval 'rs.initiate({_id:"rs0",members:[{_id:0,host:"localhost:27017"}]})'
argon doctor

main ──branch──▶ experiment ──checkout──▶ mongodb://…  ← any driver
                                              │
                     ┌── argon diff ──────────┤  document history
                     ▼                        ▼
        merge (a data PR)          or   undo / discard / rewind

# 0 · Stop source writers and DDL for the import — or: argon projects create myapp
argon import database --uri mongodb://localhost:27017 --database myapp --project myapp --source-quiesced

# 1 · Branch — instant, no copy
argon branches create experiment -p myapp

# 2 · In terminal A, capture writes with a branch actor label
argon checkout -p myapp -b experiment      # prints a connection string
argon watch    -p myapp -b experiment --actor agent:experiment

# In terminal B, write through the printed URI. Then:

# 3 · Review and merge back — a data pull request
argon diff          -p myapp -b experiment
argon merge preview -p myapp -b experiment
argon merge apply <plan-id>

# …or rewind instead of merging
argon restore reset -p myapp -b main --time 2026-07-07T09:00:00Z --backup pre-incident

claude mcp add argon -- argon mcp        # 13 tools: sandbox, diff, merge, undo, pins
# Install the v0.2.0 release wheel; PyPI currently has the older v0.1.0 API.
python3 -m pip install 'argon-agents[langgraph] @ https://github.com/argon-lab/argon-agents/releases/download/v0.2.0/argon_agents-0.2.0-py3-none-any.whl'

from argon_agents import ArgonClient, ArgonCheckpointSaver

argon = ArgonClient("http://127.0.0.1:1818")
argon.get_or_create_project("myapp")
saver = ArgonCheckpointSaver.from_sandbox(argon, "myapp")
# Compile and run your LangGraph graph with checkpointer=saver, then review:
plan = argon.merge_preview("myapp", saver.sandbox.branch)
print(plan)
# After reviewing this exact plan, apply it explicitly:
# argon.merge_apply(plan["id"])
# Or reject the run: saver.discard()

argon.create_pin("myapp", "eval-v1")                        # pin the dataset once
run = argon.sandbox_from_pin("myapp", "eval-v1")            # identical state, every eval run

Give your MongoDB a time machine. Branch without fear. ⭐