The author tested Cowork on a morning podcast from newsletters via Readwise and n8n MCP, plus video and 3D print design, noting what worked and what failed.
The author tested Cowork on a morning podcast from newsletters via Readwise and n8n MCP, plus video and 3D print design, noting what worked and what failed.
Newsletters are forwarded to Readwise. Because Cowork has no network access, the author ran n8n as an MCP server with a workflow that calls the Readwise API. Cowork fetched the previous day's articles through it and wrote a two-page Word briefing with an executive summary. It then tried to turn the briefing into a 2-3 minute MP3.
The briefing document worked and the n8n bridge ran without problems. Audio failed: with no access to a good text-to-speech model, Cowork fell back to the local Flite library, and the result was robotic and not worth listening to.
Cowork can't reach cloud text-to-speech models. The author plans to rebuild this in Claude Code, which can call the Readwise API directly and use ElevenLabs or OpenAI TTS.
Every morning at [time], gather the newsletters I received in [email or Readwise source], pick the most important stories on [my topics], and write a [number]-minute podcast-style script. Save it to [folder] as a text file, and turn it into audio if a tool is connected. Tell me which steps worked and which didn't.
Score every demo request against your ICP, update HubSpot and ping the right rep in Slack within a minute of the form submit.
Every new Zendesk ticket gets a category, priority and sentiment tag plus a draft reply in an internal note before an agent opens it.
Each month, find posts whose search clicks dropped, research what changed with Perplexity, and get a Claude-written refresh brief for each one.