A
licenseNot graded
qualityB
maintenanceEnables local discovery, reading, and messaging between running Codex and Claude Code sessions through MCP.
670 npm
MIT
tincan
OfficialRelated Servers
Alternatives to tincan
No user-submitted related servers found.
Related Servers
- AlicenseNot gradedqualityCmaintenanceLets local Claude Code and Codex sessions discover each other, read recent messages, and send questions between live or closed sessions.MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to discover and exchange text messages between existing, locally authenticated Claude Code and Codex sessions on the same machine, using installed clients without API keys.9 npm2MIT
- AlicenseNot gradedqualityAmaintenanceEnables coding agents on the same machine to discover each other as named peers and exchange messages across sessions, bridging MCP-based clients and Claude Code without a daemon.27 PyPIMIT
- AlicenseAqualityDmaintenanceEnables multiple coding agents (Claude Code, Codex, Cursor) to discover each other's sessions, search transcripts, ask questions, and handoff tasks through a shared MCP server.54 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables Claude Code sessions to discover each other as named peers and exchange instant messages across directories, machines, and Docker containers, with durable delivery for offline sessions.MIT
TDQS
A4.6/5.0
Scored across 5 tools
Disambiguation5/5
Each tool occupies a distinct slot: peers lists sessions, send_peer sends, receive_peer_messages pulls inbound content, message_log audits delivery outcomes, and reregister repairs the session's own registration. The closest pair, receive_peer_messages vs message_log, is explicitly separated (inbox content vs. a log of sends/outcomes), and the descriptions reinforce that boundary. An agent should have no trouble choosing.
Naming Consistency3/5
Names are all snake_case, but the conventions are mixed: verb_noun (send_peer, receive_peer_messages), bare noun (peers), bare verb (reregister), and noun_noun (message_log). Each name is individually readable and predictable in isolation, but there is no single pattern an agent can generalize from.
Tool Count5/5
Five tools is well-scoped for a peer-to-peer session messaging surface — send, receive, enumerate, audit, and self-repair each cover a real workflow with no filler. Nothing appears redundant or padded, and nothing obviously missing from the core loop.
Completeness4/5
The core lifecycle (discover peers, send, poll inbox, audit outcomes, recover a stale registration) is fully covered. Minor gaps: no explicit acknowledgment/consume for the pull inbox (callers must manage cursors and dedupe by ID), and inbox administration/quarantine inspection are referenced but not exposed here — though that may be deliberate scope.
Maintenance
ActivityMaintained
ResponsivenessUnresponsive