Our belief
The future is millions of agents,
orchestrated by humans.
Collaboration redesigned for agentic computing.
Storage, sync, and versioning. Rebuilt for agentic computing.
npm i -g @mod-computer/mod
Copy
Developer collaboration
Point Mod at a folder. Every file change is tracked automatically — no commits, no staging, just work.
Each agent gets its own isolated branch automatically. Work in parallel, track every change, recover any state.
Write specs collaboratively. Link requirements to code. Keep agents aligned with what you're building.
See all changes across agent sessions. Stacked diffs, grouped by feature, ready for review.
Combine work from multiple agents. Conflict detection built in. Merge when ready.
Not just files. Relationships, history, what changed — all queryable by your agents.
Build custom views into your data. Describe it, Mod generates it, share with your team.
Developer platform
mod.init()
Connect any directory. Auto-track all file changes.
mod.branch()
Create isolated branches for each agent session.
mod.query()
Find files, relationships, and history.
mod.sync()
Real-time sync across all clients.
mod.merge()
Combine branches with conflict resolution.
mod.subscribe()
Listen to changes as they happen.
import { mod } from '@anthropic/mod'
// Connect to a directory
const project = await mod.init("./my-app")
// Create a branch for this agent
const session = await project.branch("claude")
// Work on files...
await session.write("auth.ts", code)
// Query what changed
const changes = await project.query("modified today")
// Merge back to main
await session.merge()Research
We're building the infrastructure layer for a world where millions of agents work alongside humans. Our research focuses on real-time distributed systems, conflict-free data structures, and human-agent coordination at scale.
Learn more →The assumptions behind version control no longer hold when thousands of agents write code simultaneously.
Read essay → PaperConflict-free replicated data types adapted for high-throughput, multi-agent collaboration patterns.
Read paper → EssayDesign patterns for keeping humans in the loop when agents outnumber developers 100:1.
Read essay →Our belief
Collaboration redesigned for agentic computing.