Temporal MCP Server

temporal-mcp

Your model knows calculus but not what day it is. Fix that.

temporal-mcp is a tiny Model Context Protocol server that gives LLM agents a sense of time between turns. Two tools, a few hundred lines, stdlib + mcp + platformdirs. That's the whole thing.

The problem

Open a fresh chat at 11 PM. The model says "good morning." Resume a conversation three weeks later. The model picks up mid-sentence like no time passed. Ask for "today's status." Get yesterday's status. Or last Tuesday's.

LLMs don't have wall clocks. They don't know when the last user message was, whether the calendar flipped, or whether this is a fresh thread or one resumed after a long gap. Most of the time this is harmless. Sometimes it makes your agent sound like it just woke up from cryosleep.

The fix

A persistent per-thread last-seen log, exposed as two MCP tools:

• temporal_tick — call this once per user turn. Returns "it has been 14 minutes since the last message, no day rollover, timezone MDT" in a format the model can actually read.
• temporal_peek — same thing, but doesn't advance state. For when you want the gap without claiming a turn.

That's it. Time exists. Your model should know that.

Try it in 10 seconds

curl -s -X POST https://temporal-mcp.dev/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $(uuidgen)" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{
        "name":"temporal_tick",
        "arguments":{"thread_key":"try-it","tz_offset_minutes":-360,"tz_name":"MDT"}}}' \
| python3 -c 'import sys,json; print(json.load(sys.stdin)["result"]["content"][0]["text"])'

You'll see something like:

[temporal] Wed May 13, 10:42 AM MDT | fresh thread (no prior history)
{...JSON payload...}

Run it twice and the second response shows the gap. Run it tomorrow and you'll get day rollover: yes. That's the whole point.

What you get

Every tick returns a human-readable header and a JSON payload:

[temporal] Wed May 13, 9:42 AM MDT | last prompt 14m ago (Wed 9:28 AM)

{"thread_key": "mcp:abc123", "now": 1747158120.0, "prev": 1747157280.0,
 "delta_sec": 840, "day_rollover": false, "fresh_thread": false,
 "tz_name": "MDT", "tz_offset_sec": -21600, "available": true, "error": ""}

The header is for the model. The JSON is for your code, in case you want to do something interesting with day_rollover (greet differently, reload context, recompute "today's items") or with delta_sec (decay relevance, detect a resumed session, flag idle threads).

Two ways to use it

1. Hosted endpoint — claude.ai, ChatGPT, mobile

If you use claude.ai web, ChatGPT, or anything else that wants a remote MCP server, point your connector at:

https://temporal-mcp.dev/mcp

There are two ways to authenticate, depending on what your client UI exposes:

A. OAuth 2.0 (claude.ai and ChatGPT custom connectors)

Both claude.ai and ChatGPT's custom connector UIs require OAuth 2.0 with a Client ID and Client Secret. The hosted endpoint is a full OAuth provider — visit https://temporal-mcp.dev/connect and click Generate OAuth Credentials. You'll get a fresh client_id + client_secret pair, shown once. Paste them into your client's connector config. That's the entire signup.

No email, no password, no account record — the credential pair is the identity. We store only a SHA-256 of the secret, so we never see the plaintext. Generate a new pair any time you want a fresh timeline.

Claude.ai setup: Settings → Connectors → Add custom connector. URL https://temporal-mcp.dev/mcp. Paste your Client ID and Client Secret. Connect. The auto-approve flow redirects you back, claude.ai exchanges the code for a token, and you're done.

ChatGPT setup: Same idea — Settings → Connectors → Custom MCP. Same URL, same credentials.

B. Raw bearer token (Cursor, Cline, Claude Desktop, Zed, Claude Code)

If your client supports custom HTTP headers (most do), skip OAuth and just send any opaque string as a bearer token:

Authorization: Bearer <any opaque string you choose>

Pick a UUID, a passphrase, anything. We SHA-256 it before storing anything; same identity-is-the-credential property as the OAuth flow, without the dance. This is the original lowest-ceremony path and works for any client that lets you set a custom header.

Either way

No signup. No email. No PII. The hosted endpoint is free, rate-limited to 60 requests/minute per credential. If you outgrow that, self-host (see below).

2. Local stdio — Claude Desktop, Cursor, Cline, Zed, Claude Code

For desktop/IDE MCP clients, pip install the Python package and run it locally. No network round-trip, state lives on your disk, no auth needed.

pip install temporal-mcp

Python 3.9+. Linux, macOS, Windows.

Run as stdio:

temporal-mcp        # or: python -m temporal_mcp

Claude Desktop

{
  "mcpServers": {
    "temporal": {
      "command": "temporal-mcp"
    }
  }
}

Cursor / Cline / anything else that speaks MCP stdio

Same idea — point the client at the temporal-mcp command.

Self-host the remote endpoint (Cloudflare Workers)

The hosted endpoint at temporal-mcp.dev runs on Cloudflare Workers backed by D1. If you want your own instance — for privacy, scale, or to ship it as part of a larger product — the entire deploy lives in workers/:

cd workers
npm install
npx wrangler login
npx wrangler d1 create temporal_mcp           # creates the database
# Paste the printed database_id into wrangler.toml
npx wrangler d1 migrations apply temporal_mcp --remote
npx wrangler deploy

Free tier covers ~100k requests/day forever. Set REQUIRE_AUTH=true in [vars] to refuse anonymous traffic. The Worker is ~400 lines of TypeScript and has its own unit tests (workers/test/).

Tools

temporal_tick

Advance the clock for a thread and return a snapshot. Call once per user turn.

temporal_peek

Read-only. Same shape, doesn't advance state. Use it when you want the gap delta but the call isn't the canonical "one tick per user turn" event.

State

Per-thread last-seen state lives at:

Override with TEMPORAL_MCP_STATE_DIR=/some/path.

State writes are flock-safe on POSIX and atomically replaced via os.replace, so multiple agents pointing at the same state directory will not corrupt each other. (Windows falls back to an in-process lock — fine for a single MCP server, not designed for cross-process contention.)

Maintenance

python -m temporal_mcp gc        # prune threads > 30d idle
python -m temporal_mcp gc 7      # prune threads > 7d idle

Not exposed as an MCP tool on purpose — a model that can prune its own memory of "when did we last talk" will eventually do it at exactly the wrong moment. Run it from cron if you care.

Design notes (for the curious)

• Thread keying is namespaced as {client_id}:{key}. Reserve a unique client_id per surface so threads from claude.ai web don't collide with a local Cursor session sharing the same state directory.

• Failure is honest. If the state file is unreadable or the lock times out, the snapshot returns available: false with an error field and the header says gap: unknown. It does not silently lie and call it a fresh thread — a model that thinks every turn is fresh will keep saying good morning forever.

• Watchdog. tick() runs in a daemon thread with a 100 ms timeout so a stalled state read can't block your hook budget. If it times out, you get the honest-failure snapshot above.

• No HTTP transport in 0.1. Stdio only — that's what Claude Desktop, Cursor, and the other major MCP clients actually use. HTTP/SSE can land in 0.2 if there's demand.

Roadmap

• 0.2 — optional HTTP/SSE transport, conversation-ID auto-resolve from Mcp-Session-Id and friends, configurable timezone override
• 0.3 — opt-in "long gap" thresholds (return a resume: true flag past N hours) so agents can branch on resumed sessions without doing the math themselves

License

MIT. See LICENSE.

Author

Built by Garret Sutherland / MirrorEthic LLC, extracted from the temporal layer of a larger cognitive-mesh project where this primitive was load-bearing enough to deserve its own package.

---

mcp-name: io.github.MirrorEthic/temporal-mcp