Summify Platform

Build with Summify.

REST API and MCP server for ingesting content, semantic search across pods, and pulling AI-generated knowledge briefs. Available on Pro and Max plans.

1. Get an API key

Generate a key from your dashboard. Pro/Max plan required.

Open dashboard

2. Read the docs

OpenAPI 3.1 spec — import into ChatGPT, Postman, or auto-generate a client.

View spec

3. Connect via MCP

Use Summify as a Model Context Protocol server in Claude Desktop, Claude Code, or any MCP client.

MCP setup

Quickstart

End-to-end: install the SDK, enqueue your first ingestion, poll the job, and subscribe to webhooks. Code samples are paste-runnable — swap in your API key and a YouTube URL.

Install the SDK

Pick your stack. cURL users can skip this step.

# No install — just cURL + your SUMMIFY_API_KEY
export SUMMIFY_API_KEY=sk_live_...

Enqueue your first ingestion

POST a YouTube URL to /v1/ingest/youtube. Returns a job_id; the work runs async.

curl -X POST https://api.summify.io/v1/ingest/youtube \
  -H "Authorization: Bearer $SUMMIFY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ"}'

# → {"data": {"job_id": "b1a2...", "status": "queued"}}

Poll for completion

GET /v1/jobs/{id} every few seconds until status is completed or failed.

JOB_ID=b1a2...
until curl -sS https://api.summify.io/v1/jobs/$JOB_ID \
    -H "Authorization: Bearer $SUMMIFY_API_KEY" \
    | jq -e '.data.status | IN("completed","completed_partial","failed")' > /dev/null
do sleep 3; done

curl -sS https://api.summify.io/v1/jobs/$JOB_ID \
  -H "Authorization: Bearer $SUMMIFY_API_KEY" | jq .data.result

Subscribe to webhooks (optional)

Skip polling — get pushed content.ingested events as soon as a job finishes.

Register an endpoint, then verify the signature on every delivery using svix. Summify signs events with your endpoint's signing secret — treat it like an API key.

# 1. Register your endpoint
curl -X POST https://api.summify.io/v1/webhooks \
  -H "Authorization: Bearer $SUMMIFY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://yourapp.com/webhooks/summify",
    "event_types": ["content.ingested", "content.ingest_failed"]
  }'
# → {"data": {"id": "ep_...", "signing_secret": "whsec_..."}}

# 2. Fire a test event
curl -X POST https://api.summify.io/v1/webhooks/ep_.../test \
  -H "Authorization: Bearer $SUMMIFY_API_KEY"

What you can build

AI agent brains — ground your agent in curated knowledge via MCP or REST. Pod briefs become the agent's system prompt.
Ed-tech platforms — teachers create pods, students access lectures, summaries, and search.
Research workflows — ingest 500 papers, search semantically, generate synthesised briefs.
Content repurposing — one YouTube video → blog post, Twitter thread, action items, all from one ingestion.

Endpoints

Method	Path	Purpose
GET	/v1/auth/session	Verify key, get plan + credits
POST	/v1/capture	Ingest YouTube URL or voice upload
POST	/v1/text	Ingest raw text
GET	/v1/pods	List your pods
POST	/v1/pods/assign	Assign content to pod
GET	/v1/pods/{id}/content	List content in a pod
GET	/v1/pods/{id}/brief	Get AI-generated pod brief
POST	/v1/pods/{id}/search	Pod-scoped semantic search
POST	/v1/search	Global semantic search
POST	/v1/chat	Grounded chat over a pod (citations + token usage)
GET	/v1/content/{type}/{id}	Content detail (transcript + summary)
GET	/v1/styles	List summary styles
GET	/openapi.json	Full OpenAPI 3.1 spec

Coming soon

• Organization accounts with scoped keys for partner teams
• Sub-user isolation for multi-tenant apps (per-student progress, etc.)
• Async ingestion with webhook callbacks
• Quiz, flashcard, and teach-back endpoints (Learn API)
• Per-key rate limits and usage dashboards