Transcribe and summarize calls automatically

Business data & reporting · Dev, web & IT · checked 2026-10-17

Send call recordings to Deepgram’s API for transcription, have a model summarize each call, and store summaries in Notion. See the voice-assistant build for a fuller example.

Independent review. IonBin has no partnership with any tool on this page and was not paid for this page. Our evaluation is our own, based on public documentation and our research. Product names are trademarks of their owners and are used for identification only. How disclosure works.

Goal

Make every call searchable without a meeting bot.

Connects through: API · MCP

Steps

#WhoWhat happensHowStatus
1Deepgram Transcribes the recordings through its pre-recorded audio API API Confirmed · source ↗
2Your AI platform Summarizes each call AI platform Confirmed
3Notion Stores the summaries
Notion MCP can create pages and databases.
MCP Confirmed · source ↗
4You Get consent to record; spot-check summaries Human step Confirmed

Confirmed means the vendor documents it (linked). It doesn't mean IonBin has tested it. AI-platform and human steps need no source.

Human approval

Record only with consent. You spot-check summaries.

Watch out for

Keep the human approval step: Record only with consent. You spot-check summaries.

Starter prompt

Goal: Make every call searchable without a meeting bot. Tools involved: Deepgram, Notion. Your part: summarizes each call; stores the summaries. Work step by step, show me each draft or proposed change, and wait for my approval before anything is sent, published, signed, filed or changed in a live system.

Assembled from this page’s goal and steps. Paste it into your AI platform and adjust it to your situation.

Tools in this workflow

ToolRoleOfficial site
DeepgramAPI deepgram.com ↗
NotionMCP notion.com ↗

Which AI platforms can run it

PlatformStatusWhy
Claude (incl. Claude Code) To verify Notion: pairing not named in vendor docs
OpenAI Codex To verify Notion: pairing not named in vendor docs
Google Antigravity To verify Notion: pairing not named in vendor docs
ChatGPT To verify Notion: pairing not named in vendor docs
Gemini To verify Deepgram: code execution to verify (or run it in Make or n8n); Notion: pairing not named in vendor docs
Perplexity To verify Deepgram: code execution to verify (or run it in Make or n8n); Notion: pairing not named in vendor docs

Computed from the MCP and API steps only: "Confirmed" needs the tool's docs to name the platform and the platform to support the connection. Browser, upload and human steps work from any platform.

What we checked

  • Deepgram transcribes pre-recorded audio from a file or a URL through its API and SDKs (JavaScript, Python, C#, Go and Java). Streaming transcription is a separate API; for recordings you already have, use pre-recorded.
  • Pricing. Deepgram’s pricing page offers a $200 credit, then pay as you go. Per-minute prices depend on the model and on whether audio is streamed or pre-recorded, and some streaming rates are marked as limited-time promotions. Run a few calls through it and check usage before you commit to a volume.
  • Notion MCP can create pages and databases, so your AI platform can file each summary as a page in a calls database.

Recording-consent rules differ by location, and some places require every party’s consent. Tell people the call is recorded and why before you start.

Set it up

  1. Save recordings to one folder or storage bucket.
  2. Send each file to Deepgram and keep the transcript with the call date and participants.
  3. Ask your AI platform for a short summary with decisions and follow-ups, and file it in Notion.
  4. Spot-check a few summaries against the transcript each week.

See also the voice assistant build.

Sources

No affiliate links on this page. No paid placements in workflows. How IonBin makes money.