Connect Scribe & Imagine
One URL works everywhere. The first time you connect, you sign in with Google and paste your OpenAI key; after that every app you connect just asks you to press Allow.
https://scribeimagine.com/api/mcpClaude on the web and desktop
- Open Settings → Connectors and choose Add custom connector.
- Name it Scribe & Imagine, paste the URL above and press Add.
- Press Connect, sign in with Google, add your OpenAI key if asked, then Allow.
- In a chat, make sure the connector is switched on in the tools menu, then just ask.
On Team and Enterprise plans an owner may need to add the connector for the organization first. Connectors you add on the web also appear in the mobile apps.
Claude Code
Install the plugin (adds the connector plus skills that teach Claude the best workflow):
/plugin marketplace add Stronger-eCommerce/scribe-imagine-mcp
/plugin install scribe-imagine@scribe-imagineor add just the connector:
claude mcp add --transport http scribe-imagine https://scribeimagine.com/api/mcpThen run /mcp in Claude Code and choose Authenticate.
Transcription only
Want only the transcription tools (no image generation)? Use https://scribeimagine.com/api/scribe as the connector URL instead, or install the scribe@scribe-imagine plugin. Same account, same sign-in.
Other apps
Any app that supports remote MCP servers over Streamable HTTP with OAuth works: point it at the URL above.
Paying
On your account page choose your own OpenAI key (free; OpenAI bills you directly) or credits (add any amount from $5, plus a 3% + $0.30 card processing fee; each job costs OpenAI's list price plus a 20% service fee — see pricing). Every result in Claude ends with what it cost, and the account page lists each charge with the OpenAI price and the fee shown separately.
Your own OpenAI key
- Create one at platform.openai.com/api-keys and paste it on your account page. It is checked with OpenAI, encrypted, and never shown in full again.
- Images need a verified OpenAI organization (Settings → Organization → General → Verify).
- Set a monthly budget in OpenAI → Settings → Limits if you want a hard cap.
- A restricted key works if it allows Models (read), Images and Audio.
Things to ask
- “Make a flat illustration of a fox reading a newspaper, square, low quality first.”
- “Now make the background deep blue and add the title ‘Morning Edition’.”
- “Put our logo from https://… on a coffee mug mockup.”
- “Transcribe this call with speaker labels — it’s our weekly sync with Dana and Sam: <Drive link>.”
- “Speaker A is Dana, B is Sam. Now summarise the decisions and action items.”
- “Make subtitles for this video: <link>.”
No link handy? Upload the file at /transcribe and then ask Claude to “get my latest transcript”.
Tools
generate_image | Make pictures from a description (quality low → max, square/landscape/portrait, PNG/JPEG/WebP, transparent backgrounds). Runs in the background and returns a job id. |
edit_image | Change a generated image by id, or edit/combine your own pictures from public links. |
image_status | Collect a finished image job (the picture plus a shareable link). |
list_images / delete_image | See what's stored; delete pictures and break their links. |
transcribe | Transcribe a recording from a link (Drive, Dropbox, OneDrive or direct), up to 4 hours: best or fast quality, speaker labels, timestamps and subtitles, with context, participants and keywords for accuracy. |
transcription_status | Check a long transcription and get the result. |
get_transcript | Read a saved transcript a page at a time, as text, .srt, .vtt or JSON. |
rename_speakers | Replace Speaker A/B with real names. |
list_transcripts / delete_transcript | Your saved transcripts, including uploads from the website. |
account_status | Which account is connected, whether a key is saved, and useful links. |
Limits
- Recordings: any audio or video format, up to 4 hours and 450 MB. Links must be publicly downloadable.
- Speaker labels stay consistent for up to four voices across a long recording.
- The newest 40 images (up to 150 MB) are kept per account; older ones are dropped.
- Uploaded recordings are deleted as soon as they are transcribed.
Support
Questions or problems: email jay@strongerecommerce.com. Scribe & Imagine is free; if it helps you, .