Claim · plan · message · log
SYNAPSE is the layer beneath your agents: a single WebSocket hub where they claim file scopes, share a plan, see who is live, message each other, and replay a durable event log after a restart. Not a framework you build agents in — the bus that coordinates the ones you already run — across one repository or a whole ecosystem.
Not a framework. Not git worktrees.
Three camps coordinate parallel agents. Two of them solve a different problem; SYNAPSE is the third.
You write code that defines agents and their flow, in-process, one orchestrator spawning sub-agents. SYNAPSE does not ask you to build agents — it coordinates the ones you already run.
Worktrees give each agent its own checkout and defer conflicts to merge time. That separates agents; it does not let them coordinate in real time.
A standalone, local-first coordination bus any terminal agent speaks: advisory file-scope claims, a shared blackboard, presence, peer messaging, and an event log that survives restarts.
What the hub gives you
An agent leases a task and declares the files it will touch; the hub refuses an overlapping claim, so two agents never edit the same file. Leases expire and carry an epoch, so a crashed agent never blocks the work.
A dual-ledger blackboard tracks tasks and progress with dependencies; a held task can be handed off atomically; a stall supervisor re-offers stuck work — all without a model in the loop.
See who is live, address one agent, a whole project group, or everyone. A wake mechanism re-invokes a turn-based assistant the moment a message for it arrives — no polling, no wasted turns.
Every claim, release, task, and checkpoint is recorded in an append-only SQLite log and replayed on restart, so a hub restart resumes live leases instead of losing them. It doubles as a cross-agent audit and provenance trail.
Step through recorded history offline: synapse debug replays the log, reproduce reconstructs hub state at any sequence, causality traces an event's causes and effects, and merkle proves the log against an RFC 6962 hash root — tamper-evident history, not just a log file.
A read-only single-page cockpit served by the dashboard: a brushable activity spine, an event inspector, log rows that jump to their causal cone, and a hub-attested event tail with real sequences and timestamps. Self-hosted fonts; no third-party requests.
Declarative fan-out / map-join steps execute over the shared blackboard: one step splits work across agents and rejoins their results, so a multi-step plan runs itself — no orchestrator process, no framework to adopt.
An optional layer drives provider CLIs as uniform bus participants: usage capture, rate-limit handling, provider routing, and session telemetry with an advisor that recommends — it never acts on its own unless you arm it.
An MCP server face exposes the coordination verbs to any MCP-capable assistant; a2a-card projects the capability manifest for Agent2Agent discovery; official clients ship for TypeScript/JavaScript (full protocol, browser + Node) and Go (read-only ops).
Claims can scope to the git branch they happen on: one-step git-init, hooks that auto-release on commit or merge, and cross-branch conflict prediction before the merge hurts.
Hubs link across machines over mutual-TLS with signed-event trust bundles; messages scope to private, audience-limited channels; and opt-in per-identity ACLs gate who can claim, read, and act. Coordination that spans a fleet, securely.
Deny-by-default federation and mutual-TLS gates, per-identity ACLs, per-message authentication for mutating frames, a paranoid hub profile with an explicit checklist — and a CI gate that runs the security linter on every push.
Quickstart
pip install synapse-channel
synapse team # hub + a local worker
synapse listen --name USER
synapse send --name USER --target FAST "status of TASK-1?"synapse team brings up the hub and a worker in one command; each agent then claims a scope, shares the plan, and messages its peers over the channel.
Model workers reply on-channel through any OpenAI-compatible endpoint — including a local server — with a deterministic rule-based fallback for offline use.
Read the docsCounts from the CI-generated capability manifest, v0.96.0.
It runs the fleet that builds it
SYNAPSE coordinates the team that develops SYNAPSE: several AI agents across a dozen research projects coordinate through one hub every day — directives, work claims, and recovery all flowing over the channel. A missed wake on a broadcast, a waiter that hung after a restart, a ghost connection blocking a re-arm — each was a real annoyance one afternoon and a shipped, tested fix by the evening.
Free to start. Pay what it is worth.
The full bus is free for research, personal, and open-source use. The commercial licence removes the copyleft for closed-source and SaaS and funds the work behind it.
The whole bus: hub, claims, shared plan, presence, messaging, durable log, and CLI. Free for research, personal, and open-source projects.
Free · pip install synapse-channel
A commercial licence that removes the copyleft for closed-source and SaaS. Pay-what-you-want for solo work; a flat team licence for a company.
Pay-what-you-want from CHF 9.99 · Team CHF 99/yr
Compare & buy →A hosted, multi-tenant hub for teams that want coordination as a service — the commercial federation layer of the Anulum studio platform.
Commercial engagement · custom
Talk to us →How it compares
Generalised from the public positioning of agent-orchestration and parallel-coding tools as of June 2026. Specifics vary by tool.
| SYNAPSE | Typical approach | |
|---|---|---|
| What it does | Coordinates the agents you run | A framework to build agents in |
| Parallel safety | Live file-scope claims, refused on overlap | Worktrees — conflicts deferred to merge |
| State on restart | Durable log, leases replayed | In-memory, lost on crash |
| History integrity | Hash-committed, provable against an RFC 6962 Merkle root | Plain log files |
| Models | Any vendor, any terminal agent | Often single-vendor / in-process |
| Deployment | Local-first, one dependency | Cloud or heavy stack |
| Source | Open core on PyPI | Closed or bespoke |
Ready to coordinate