Install Motif as an MCP server in any host that speaks the Model Context Protocol — Claude Code, Cursor, OpenClaw, Windsurf, Continue, Raycast, and more. Your agent gets voice-trained content repurposing across 6 platforms as native tool calls, no HTTP client required.
Both work. The REST API (/developers) is right when you're writing your own client. The MCP server is right when you want an AI agent to drive the pipeline — because the agent sees structured tool schemas it can reason about, and doesn't need any HTTP plumbing prompt-engineered into context.
Same capabilities, different protocol. Same rate limits. Same MOTIF_API_KEY. Ship which one fits — or both.
Grab an API key at Settings → API Keys (requires an active plan), then pick your host.
Anthropic's agentic CLI. One command, ready to use in the next session.
claude mcp add motif npx @motif/mcp-server \
--env MOTIF_API_KEY=motif_YOUR_KEY_HEREDrop into .cursor/mcp.json and restart Cursor. Tools appear under the Agent panel.
// .cursor/mcp.json
{
"mcpServers": {
"motif": {
"command": "npx",
"args": ["-y", "@motif/mcp-server"],
"env": {
"MOTIF_API_KEY": "motif_YOUR_KEY_HERE"
}
}
}
}Register as an OpenClaw skill. Your local agent then has the full Motif toolset across all 20+ messaging channels OpenClaw supports.
// skills/motif.js
module.exports = {
name: "motif",
mcp: {
command: "npx",
args: ["-y", "@motif/mcp-server"],
env: { MOTIF_API_KEY: process.env.MOTIF_API_KEY },
},
};The MCP spec is standardized — any compliant host can load Motif with the same config. Check your host's docs for where to put the config file.
// Any MCP-compatible host
{
"mcpServers": {
"motif": {
"command": "npx",
"args": ["-y", "@motif/mcp-server"],
"env": { "MOTIF_API_KEY": "motif_YOUR_KEY_HERE" }
}
}
}Five tools cover the full content-repurposing surface. All return structured JSON your agent can reason about without re-prompting.
| Tool | Purpose |
|---|---|
| motif.repurpose | Run the full content-repurposing pipeline. Input text or URL, get 6 platform-native posts in your trained voice. |
| motif.generate | Generate a single platform-native post from an idea + voice profile. Use when you want tight control over individual outputs. |
| motif.extract_ideas | Pull 8–15 ranked shareable ideas out of a long-form source. Feeds into motif.generate for per-idea posts. |
| motif.voices | List your voice profiles, or train a new one from writing samples. |
| motif.jobs | Check async job status. Pipelines run in the background; use this to poll for completion. |
Authentication uses the same MOTIF_API_KEY as the REST API. Set it in the host's MCP env block; the MCP server passes it on every tool call. Keys are generated at /settings/api-keys and revocable at any time.
Rate limit: 60 requests per minute per key. Rate-limited responses return HTTP 429 with a Retry-After header. Plan limits (10 or 30 pipeline runs/month on Starter or Pro) apply across both MCP and REST.
Once installed, prompt your agent in plain language:
# In any MCP host's chat:
"Use motif.repurpose on this URL and save the
outputs to a file: https://myblog.com/latest-post"
# The agent will:
# 1. Call motif.repurpose({ url: "..." })
# 2. Receive a jobId + polling URL
# 3. Poll motif.jobs until status is "completed"
# 4. Fetch the 6-platform outputs
# 5. Write them to disk (or anywhere else you asked)Expect ~90 seconds for a typical repurpose job end-to-end — most of that is LLM generation across 6 platforms.
Bugs, questions, or a host you'd like Motif's MCP server to be tested on? Email declan.paul97@gmail.com. Reply within a day.
MCP server + REST API. $24/mo (annual) to get started. 7-day money-back guarantee.
From $24/mo (annual) or $29/mo · 7-day money-back guarantee · Cancel anytime