# Rhylthyme MCP server

[![Claude Connectors Directory](https://img.shields.io/badge/Claude%20Directory-Rhylthyme-D97757)](https://claude.ai/directory/rhylthyme)
[![Listed on mcpservers.org](https://mcpservers.org/badge.svg)](https://mcpservers.org/servers/rhylthyme/rhylthyme-mcp)

A [Model Context Protocol](https://modelcontextprotocol.io/) server for
schedules that a person executes: cooking several dishes so they finish
together, running a bench protocol with overlapping incubations, calling
an event, running a workout. An agent describes the process, the server
validates and analyzes the resulting program, and publishes it as a live
timeline on [rhylthyme.com](https://www.rhylthyme.com) that the person
follows on their phone.

This repository is the source of the server that runs at
`mcp.rhylthyme.com`. It is the same code as `mcp-api/` in the
rhylthyme-server application, copied here with the static assets it
needs so it can be read, tested and self-hosted.

## Try it: ask for a workout

Connect the server (in claude.ai, Claude Desktop or mobile, add
[Rhylthyme from the Claude Connectors Directory](https://claude.ai/directory/rhylthyme);
in Claude Code, `/plugin marketplace add rhylthyme/rhylthyme-mcp`
then `/plugin install rhylthyme@rhylthyme`; anywhere else, add
`https://mcp.rhylthyme.com/gym/mcp` as a connector), then say what you want in
plain words:

> Two of us in a garage gym: one kettlebell, one pull-up bar, one jump rope.
> Thirty minutes, two rounds, and nobody stands around waiting for equipment.
> Make it a Rhylthyme timeline.

You write no JSON. The assistant does, and the server keeps it honest. With
that prompt Claude made three tool calls:

| Call | What came back |
|---|---|
| `validate_program` | `✅ Program is valid — 2 tracks, 20 steps, 30m makespan.` |
| `analyze_schedule` | `Resource conflicts: none.` Each station has capacity 1, so this is the proof that Alex and Sam never want the kettlebell at the same moment |
| `visualize_schedule` | a link, [gym.rhylthyme.com?share=4995ef1b88744e80](https://gym.rhylthyme.com?share=4995ef1b88744e80), with rest timers and interval beeps, and a text chart for the chat: |

```
                0                       15m                    30m
Alex           │░Warm-…░▒S…▒░Pu…░▒Ju…▒░░░▒Sw…▒░Pu…░▒Ju…▒░░░▒St…▒│
Sam            │░Warm-…░▒P…▒░Ju…░▒Sw…▒░░░▒Pu…▒░Ju…░▒Sw…▒░░░▒St…▒│
```

Ask "show me a picture" and `preview_timeline` returns an image in the chat.
For one coloured by station, which makes the rotation obvious, save the
program the assistant wrote (ours is [`examples/recipe/garage-circuit.json`](examples/recipe/garage-circuit.json)) and run:

```bash
npx -y github:rhylthyme/rhylthyme-timeline garage-circuit.json -o garage-circuit.png \
  --style web --palette tableau --color-by task
```

![Garage circuit for two, coloured by station: the kettlebell, the bar and the rope are never double-booked](https://raw.githubusercontent.com/rhylthyme/rhylthyme-mcp/main/examples/recipe/garage-circuit.png)

Then push on it: "add a third person", "we only have 20 minutes", "swap the
rope for burpees". The assistant edits the program and the server checks it
again.

## This repository and rhylthyme

Two repositories do related things and are easy to confuse. This one is the
**server an AI assistant talks to**. [rhylthyme](https://github.com/rhylthyme/rhylthyme)
(formerly rhylthyme-cli-runner) is the **command a person types**.

| | [rhylthyme-mcp](https://github.com/rhylthyme/rhylthyme-mcp) | [rhylthyme](https://github.com/rhylthyme/rhylthyme) |
|---|---|---|
| What it is | The MCP **server**: the tools an AI assistant calls | A command-line **program**: the `rhylthyme` command |
| Who uses it | Claude, ChatGPT, Cursor or any MCP client, on a person's behalf | A person at a terminal, a script, or CI |
| Where it runs | Hosted at `mcp.rhylthyme.com`; nothing to install | On your machine: `pip install rhylthyme` |
| Language | JavaScript (Node 20+) | Python 3.12+ |
| Input | A program the assistant builds in conversation | A program file on disk (JSON or YAML) |
| Validate a program | `validate_program` | `rhylthyme validate` (works offline) |
| Timing, conflicts, deadlines | `analyze_schedule` | `rhylthyme analyze` (asks the server) |
| Publish a live timeline | `visualize_schedule` | `rhylthyme publish` (asks the server) |
| Run a schedule with timers | no: it hands back a link to the web timeline | `rhylthyme run`, an interactive terminal runner |
| Recorded runs, calibration | reads runs saved to an account | records runs locally; `rhylthyme runs`, `rhylthyme calibrate` |
| Catalog search, imports, account library | yes | no |
| Also in the repository | the `rhylthyme-mcp` PyPI package (a stdio bridge to the hosted server), the Claude plugin marketplace | the Claude skill's source, the prompt-evaluation harness and its results, `rhylthyme mcp-test` |

How they fit together: the command-line tool is one of this server's clients.
`rhylthyme analyze`, `publish`, `generate` and `mcp-test` are MCP calls to
`mcp.rhylthyme.com`; `rhylthyme validate`, `run`, `runs` and `calibrate` never
touch the network. Each has its own validator for the same program schema
(JavaScript here, Python there), so a program is checked again when it is
published.

Use this repository to connect an assistant, to read or self-host the server,
or to change a tool. Use rhylthyme if you have a program file and a
terminal, want timers in the terminal, or keep run records.

Two names to keep apart: `rhylthyme-mcp` on PyPI is this server's stdio bridge
(command `rhylthyme-mcp`, source in [`python/`](python)); `rhylthyme`
on PyPI is the command-line tool (command `rhylthyme`; it was published as
`rhylthyme-cli-runner` before 0.5.0). The program format
itself is defined in [rhylthyme-spec](https://github.com/rhylthyme/rhylthyme-spec),
with examples in [rhylthyme-examples](https://github.com/rhylthyme/rhylthyme-examples).

## Endpoints

| URL | Server name | Adds |
|---|---|---|
| `https://mcp.rhylthyme.com/mcp` | `rhylthyme-mcp` | generic scheduler |
| `https://mcp.rhylthyme.com/kitchen/mcp` | `rhylthyme-kitchen-mcp` | `cook_recipe`, `whats_for_dinner` |
| `https://mcp.rhylthyme.com/lab/mcp` | `rhylthyme-lab-mcp` | `run_protocol`, `random_protocol`, Benchling import |
| `https://mcp.rhylthyme.com/events/mcp` | `rhylthyme-events-mcp` | `plan_event`, `random_event_template` |
| `https://mcp.rhylthyme.com/gym/mcp` | `rhylthyme-gym-mcp` | `start_workout`, `surprise_workout` |

Transport: Streamable HTTP, stateless. No sign-in is needed for the public
catalog or the pure tools; account tools use the person's Rhylthyme account
through OAuth 2.1, or a pasted token from the `login` tool in clients without
OAuth. Server `instructions` describing the workflow are sent at
`initialize`.

## Quickstart: timelines from the command line

The `rhylthyme` CLI can drive this server directly: describe what you
need in plain language and get back a live timeline URL, a program file,
or both.

```bash
pip install rhylthyme                        # Python 3.12+
rhylthyme login                              # opens rhylthyme.com in your browser
rhylthyme generate "roast chicken, potatoes and green beans for 6" \
    -e kitchen --by 19:00 --with "one oven, four burners, one cook"
```

`login` signs you in through the browser and hands the session back to a
one-shot listener on `127.0.0.1`. It is stored in
`~/.config/rhylthyme/credentials.json` (mode 0600) and renews itself, so
you only log in once. `generate` then calls two tools on this server:

1. **`import_text`** on the endpoint for `-e` (`/kitchen/mcp`,
   `/lab/mcp`, …) turns the request into a validated multi-track program.
   It runs four model turns server-side, which is why it needs a sign-in;
   it takes 20–60 seconds and is capped per day.
2. **`visualize_schedule`** publishes that program and returns the
   live-timeline URL on the matching subdomain.

It prints the ASCII Gantt, the itinerary and the URL. More examples:

```bash
# A lab protocol from a file; save the program and run it in the terminal
rhylthyme generate -e lab -f western_blot.txt -o blot.json --run

# Pipe a run sheet in; print only the URL
pbpaste | rhylthyme generate -e events --by "doors at 18:30" -q

# Program JSON only, no published timeline; machine-readable output
rhylthyme generate -e gym "45 minute upper-body circuit, two people, one bench" --no-publish --json
```

| Flag | Meaning |
|---|---|
| `-e, --env` | `generic` (default), `kitchen`, `lab`, `events` or `gym`. Picks the endpoint and the timeline site. |
| `--by` | When everything must be finished: `19:00`, `dinner at 7pm`. |
| `--with` | Equipment and people limits in your own words. |
| `-f, --file` | Read the request or source text from a file, or `-` for stdin. |
| `-o, --output` | Save the program JSON. |
| `--run` | Run the program in the terminal runner afterwards. |
| `--open` | Open the live timeline in a browser. |
| `--no-publish`, `--json`, `-q` | Skip publishing; print JSON; print only the URL. |

To check a server (this one, or your own deployment) end to end:

```bash
rhylthyme mcp-test                                   # all five hosted endpoints, read-only
rhylthyme mcp-test --url http://localhost:3000/mcp -e generic --publish
```

`rhylthyme whoami` shows the stored sign-in; `rhylthyme logout` forgets it.
On a machine without a browser, set `RHYLTHYME_TOKEN` to an access token
from <https://www.rhylthyme.com/mcp/auth> (it lasts about an hour), or
run `rhylthyme login --token <token>`. `RHYLTHYME_MCP_URL` points the CLI
at a self-hosted server.

## Connect

**Claude Code**

```bash
claude mcp add --transport http rhylthyme https://mcp.rhylthyme.com/mcp
claude mcp add --transport http rhylthyme-kitchen https://mcp.rhylthyme.com/kitchen/mcp
```

**Claude Desktop / claude.ai**: Settings → Connectors → *Add custom
connector* → paste one of the URLs above. Public tools work immediately;
run `login` only to save to your own account.

**Claude Code plugin** (the hosted server plus a skill that teaches Claude to
author, validate and analyze schedules):

```
/plugin marketplace add rhylthyme/rhylthyme-mcp
/plugin install rhylthyme@rhylthyme
```

**ChatGPT**: Settings → Apps & Connectors → Advanced settings → turn on
*Developer mode*, then *Create* a connector with one of the URLs above as the
MCP server URL. Without a connector ChatGPT cannot call these tools and falls
back to browsing the website.

**Cursor** (`.cursor/mcp.json`):

```json
{ "mcpServers": { "rhylthyme": { "url": "https://mcp.rhylthyme.com/kitchen/mcp" } } }
```

**Claude API** (MCP connector, one request):

```python
import anthropic
client = anthropic.Anthropic()
resp = client.beta.messages.create(
    model="claude-opus-5", max_tokens=16000,
    betas=["mcp-client-2025-11-20"],
    mcp_servers=[{"type": "url", "url": "https://mcp.rhylthyme.com/kitchen/mcp", "name": "rhylthyme"}],
    tools=[{"type": "mcp_toolset", "mcp_server_name": "rhylthyme"}],
    messages=[{"role": "user", "content": "Plan Thanksgiving for 8 with one oven, dinner at 6pm."}],
)
```

**Reviewing an import.** `review_program` (an account tool) has a model read an
imported program against its source and return findings: wrong durations,
dropped steps, bad ordering, a total that disagrees with the source. Call it
after `import_from_source` or `import_text`, apply what it says, and validate
again.

**Clients that can only launch a command**: `pip install rhylthyme-mcp` gives a
`rhylthyme-mcp` command, a stdio bridge that passes every request through to
the hosted server (source in [`python/`](python)):

```json
{ "mcpServers": { "rhylthyme": { "command": "rhylthyme-mcp", "args": ["kitchen"] } } }
```

**No MCP client at all** (an agent with a shell, a script): the server is
stateless, so one POST works with no handshake and no account.

```bash
curl -s https://mcp.rhylthyme.com/mcp \
  -H 'Content-Type: application/json' -H 'Accept: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call",
       "params":{"name":"validate_program","arguments":{"program":{"programId":"x","name":"x","tracks":[]}}}}'
```

`visualize_schedule` called the same way returns the live-timeline URL in
`result.structuredContent.url`. Or use the CLI in the quickstart above.

## Tools

| Tool | Annotations | What it does |
|---|---|---|
| `validate_program` | read-only, pure | Structural and scheduling checks (duplicate/missing ids, dangling references, cycles, within-track overlaps, tasks without a resource constraint, unparseable durations, choice references). Every finding carries a `code`, `message` and `fix` hint. Structured output. |
| `analyze_schedule` | read-only, pure | Resolved start/end per step, makespan, the dependency chain that determines it, resource-conflict windows, per-track slack, peak concurrency vs. declared actors. Pass `finishAt` or `startAt` (ISO 8601) for wall-clock start times. Structured output. |
| `visualize_schedule` | publishes | Validates (refuses invalid programs unless `allowInvalid`), creates a share record, returns a markdown preview (cover image, equipment, ingredients, ASCII Gantt, itinerary, schedule check), an inline PNG of the timeline and the live URL. Structured output. |
| `preview_timeline` | publishes a share record | PNG of the timeline only, no prose. |
| `search_public_recipes` | read-only | Keyword search over the public catalog; `environment` selects kitchen (default), laboratory, event or gym. Structured output. |
| `load_public_recipe` | read-only | Full summary and live URL for one catalog entry. |
| `cook_recipe` / `run_protocol` / `plan_event` / `start_workout` | read-only | One-shot on each vertical: top catalog match → live URL. |
| `whats_for_dinner` / `random_protocol` / `random_event_template` / `surprise_workout` | read-only | Random catalog pick on each vertical. |
| `import_from_source` | read-only | Spoonacular, TheMealDB, protocols.io, Cooklang, Opentrons Protocol API scripts, Benchling → validated program JSON. `search` needs no token; `import` and `random` need the user's token. |
| `create_environment` | pure | Equipment limits and actor types → environment JSON. |
| `login` | — | Returns the sign-in URL, then verifies a pasted token. |
| `list_my_programs`, `load_program`, `save_program` | account | The user's own library; `save_program` validates first. |
| `get_renderer_source` | read-only | Source of the Apache-2.0 timeline renderer, for HTML artifacts whose sandbox blocks external scripts. |

Annotations (`readOnlyHint`, `destructiveHint`, `idempotentHint`,
`openWorldHint`) are set on every tool; four tools also declare an
`outputSchema` and return `structuredContent`. Failures set `isError`
and say what to do next.

## Resources and prompt

- `rhylthyme://schema/program`: the program JSON Schema (0.3.0-alpha)
- `rhylthyme://guide/authoring`: one-page authoring rules, trigger vocabulary, how to make tracks finish together, repeated work (`replicates`, per-instance `instances: "each"` chains, `"all"` barriers and `maxInFlight`), choice branching
- `rhylthyme://guide/tools`: the long form of every tool description. The tool list itself is kept to about 3,600 tokens so it is cheap to keep connected
- `rhylthyme://examples/{breakfast_schedule, lab_experiment, stir_fry_with_choice, hiit_cardio_workout, corporate_presentation, cookies_three_trays}`: complete valid programs (`cookies_three_trays` is the 0.3.0-alpha per-instance / in-flight worked example)
- prompt `plan_schedule(goal, finishAt?, constraints?)`: walks the model through search → build → validate → analyze → visualize, and names `replicates` / `instances` / `maxInFlight` in its constraints step

## What a program looks like

```json
{
  "schemaVersion": "0.1.0",
  "programId": "eggs-and-toast",
  "name": "Eggs and toast",
  "tracks": [
    { "trackId": "eggs", "name": "Eggs", "steps": [
      { "stepId": "whisk", "name": "Whisk", "task": "prep",
        "duration": { "type": "fixed", "seconds": 60 },
        "startTrigger": { "type": "programStart" } },
      { "stepId": "cook", "name": "Cook", "task": "stove",
        "duration": { "type": "variable", "minSeconds": 120, "maxSeconds": 240, "defaultSeconds": 180 },
        "startTrigger": { "type": "afterStep", "stepId": "whisk" } } ] },
    { "trackId": "toast", "name": "Toast", "steps": [
      { "stepId": "toast", "name": "Toast", "task": "toaster",
        "duration": { "type": "fixed", "seconds": 180 },
        "startTrigger": { "type": "afterStep", "stepId": "cook", "event": "start", "offsetSeconds": 60 } } ] }
  ],
  "resourceConstraints": [
    { "task": "prep", "maxConcurrent": 1 },
    { "task": "stove", "maxConcurrent": 2 },
    { "task": "toaster", "maxConcurrent": 1 }
  ]
}
```

Steps in one track run sequentially; parallel work goes in separate
tracks; every `task` needs a resource constraint; durations and offsets
take seconds or strings like `"5m"`. Triggers: `programStart`,
`programStartOffset`, `afterStep` (end or `event: "start"`, signed offset),
`afterStepWithBuffer`, `manual`, `onAbort`, or `{logic: all|any, triggers}`.
Durations: `fixed`, `variable` (ended early by the executor), `indefinite`
(ended by the executor).

## Self-hosting

```bash
git clone https://github.com/rhylthyme/rhylthyme-mcp
cd rhylthyme-mcp
npm install            # Node 20 or newer
npm test                 # SDK in-memory + HTTP entry-point tests, no network
PORT=3000 npm start      # http://localhost:3000/mcp and the four vertical paths
npm run stdio            # the same server over stdio (add kitchen, lab, events or gym)
```

Stdio clients run it as a local process, for example in Claude Desktop:
`{"command": "node", "args": ["/path/to/rhylthyme-mcp/src/stdio.js"]}`.

Docker: `docker build -t rhylthyme-mcp . && docker run -p 3000:3000 rhylthyme-mcp`.
Vercel: `vercel` in the repository root; `vercel.json` rewrites the endpoint
paths to the function.

What self-hosting does and does not give you: validation, timing analysis,
the renderer, resources and prompts run in your process. Catalog search,
sharing (`visualize_schedule`), imports and account tools call the public
API at `https://www.rhylthyme.com` (`API_BASE` in `mcp-api/index.js`), so
those still depend on the hosted service. The PNG preview route
`/api/og/timeline.png` additionally needs `SUPABASE_URL` and
`SUPABASE_ANON_KEY` for read access to shared programs; nothing else does.

## Layout

- `mcp-api/index.js`: tool, resource and prompt registration; vertical detection; OG-image route; Vercel handler
- `mcp-api/schedule.js`: validator and analyzer (pure)
- `static/js/timeline-render.js`: timing engine and SVG Gantt (also published as [`@rhylthyme/timeline`](https://github.com/rhylthyme/rhylthyme-timeline))
- `static/schema/`, `static/examples/`: the resources
- `src/index.js`: standalone HTTP runner
- `src/stdio.js`: standalone stdio runner (same tools, resources and prompts)
- `mcp-api/server.json`: MCP registry manifest
- `.claude-plugin/marketplace.json`, `plugins/rhylthyme/`: the Claude plugin marketplace and plugin (`claude plugin validate .`); the skill in it is a checked copy of `skills/rhylthyme` in [rhylthyme](https://github.com/rhylthyme/rhylthyme)
- `python/`: the `rhylthyme-mcp` PyPI package, a stdio bridge to the hosted server (`cd python && PYTHONPATH=src pytest tests`)

## Known limitations

- `login` hands the user a URL and expects a pasted, short-lived access token that then travels as a tool argument; it is outside MCP's OAuth 2.1 flow. Hosts that require OAuth for authenticated servers can still use every public tool.
- `search_public_recipes` searches one environment at a time; the generic endpoint defaults to kitchen.
- Tool annotations are self-declared hints; hosts may ignore them.

## License

Apache-2.0, like the rest of Rhylthyme. (This repository was MIT until
September 2026.)