Medium
Long-form publishing.
Verdict
Common use cases
- Draft and publish stories from chat
- Pull analytics for recent posts
- Schedule editorial calendar reviews
- Coordinate multi-author publication workflows
- Repurpose internal docs as Medium posts
Integration
- Vendor
- Medium
- Category
- communication
- Auth
- OAUTH2
- Composio slug
medium
Tools
Setup
Setup guide
- 11. In Switchy, open your workspace settings and navigate to the Integrations tab. 2. Search for Medium and click Connect. 3. You'll be redirected to Medium's OAuth consent screen — sign in with the account that has author or editor permissions for your publication. 4. Grant the requested scopes: basicProfile (to identify your account), listPublications (to see publications you can post to), publishPost (to create and update stories), and uploadImage (to attach media). 5. Medium redirects you back to Switchy with a success confirmation. 6. Open any Space, type '@Medium list my publications' and send — if you see your publication names in the response, the connection is live. 7. To publish a draft, @mention Medium again with your story title, body text, and target publication, then ask it to create the post.
What teammates see: by default, memories from Medium are scoped to the Space (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Spaces have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
List My Publications
@Medium show me all the publications I can post to, including their names and IDs.Open in a Space →
Draft a New Post
@Medium create a draft post titled 'Five lessons from our Q1 sprint' with this outline: [paste outline]. Tag it 'product management' and 'agile'.Open in a Space →
Publish to Publication
@Medium publish this post to our company publication: [paste markdown content]. Set the canonical URL to our blog and mark it as published immediately.Open in a Space →
Retrieve Story Stats
@Medium get the stats for my last five published stories — views, reads, and read ratio for each.Open in a Space →
Update Existing Story
@Medium update the story with ID [story-id] — replace the second paragraph with this revised text: [paste new paragraph].Open in a Space →
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Medium fetch my last 5 published stories and show me their view counts and read ratios
Retrieved 5 stories from your Medium account. "Building Async Pipelines in Python" (Mar 15) has 2,341 views with a 47% read ratio. "Why We Migrated to Postgres" (Mar 8) has 1,829 views, 52% read ratio. "Debugging Production Memory Leaks" (Feb 28) has 4,102 views, 61% read ratio. "Team Rituals That Actually Work" (Feb 20) has 987 views, 38% read ratio. "Introduction to gRPC" (Feb 12) has 3,456 views, 55% read ratio. Read ratio indicates the percentage of viewers who scrolled past the preview.
This example demonstrates read access to your publication analytics. Requires OAuth scope for user statistics. Medium's API returns aggregated metrics, not real-time data — expect a 1-2 hour delay on fresh posts. View counts reflect total impressions across Medium's platform and external referrers.
@Medium draft a new story titled "Lessons from Q1 Incident Response" with the outline from our retro doc, then save it as a draft
Created draft story on Medium with the title "Lessons from Q1 Incident Response". The post includes sections on detection lag, communication breakdowns, and tooling gaps drawn from your retro outline. Draft is saved to your account and visible only to you. You can continue editing in Medium's web editor or publish directly when ready. Draft URL: medium.com/@yourhandle/draft-abc123.
This example shows content creation via the MCP. The AI structures your outline into Medium's format but does not auto-publish — you retain editorial control. Requires OAuth write scope. Medium enforces rate limits on draft creation (roughly 10 per hour). Images and embeds must be added manually in the Medium editor.
@Medium analyze engagement patterns across my last 20 posts and suggest which topics resonate most with my audience
Analyzed 20 recent stories. Posts tagged "incident response" and "observability" average 3,200 views and 58% read ratio. "Team culture" posts average 1,400 views, 42% read ratio. "Tutorial" posts average 2,800 views but 67% read ratio — smaller reach, higher engagement. Your audience skips listicles (avg 31% read ratio) but finishes technical deep-dives. Recommendation: prioritize hands-on tutorials and postmortem breakdowns over high-level commentary.
This example pairs Medium's analytics data with AI reasoning to surface patterns you might miss. The MCP fetches raw metrics; the AI interprets trends. Accuracy depends on sample size — 20 posts gives directional insight, not statistical certainty. Medium's API does not expose reader demographics or referral sources, so recommendations focus on content type and engagement depth.
Use-case deep-dives
When Medium MCP fits a regular content cadence
A 6-person engineering team publishes a weekly technical blog to Medium, rotating authorship across the team. The Medium MCP works well here because OAuth2 lets each engineer authenticate once, then the AI can draft posts in their voice, pull in code snippets from Slack threads or GitHub discussions, and format everything to Medium's spec before human review. The workflow stays in Switchy instead of context-switching to Medium's editor. The trade-off: if your team publishes less than twice a month, the OAuth setup overhead probably isn't worth it—just paste into Medium directly. If you're publishing daily or need multi-author coordination with tight deadlines, this MCP keeps the process moving without bottlenecking on one person's Medium login.
Medium MCP for marketing teams shipping narratives
A 3-person marketing team at a B2B SaaS company writes customer case studies and posts them to Medium as part of their content strategy. The Medium MCP lets the AI pull interview notes from Google Docs, combine them with product usage data from internal dashboards, and generate a structured draft that matches Medium's formatting conventions. OAuth2 means each marketer can review and publish under their own byline without sharing credentials. The boundary: if your case studies require heavy visual design or custom embeds that Medium doesn't support natively, you'll still need to finish in Medium's editor or move to a different platform. For text-heavy narratives with standard images, this MCP handles 80% of the formatting work and keeps the team in one workspace.
When a solo founder should skip the Medium MCP
A solo founder wants to publish monthly thought leadership pieces to Medium to build their personal brand. The Medium MCP is overkill here. OAuth2 setup, AI-assisted drafting, and workspace coordination all assume a team workflow with multiple contributors or a high publishing frequency. A single person writing once a month will spend more time configuring the MCP than they'd save by automating the publish step. The better play: draft in Switchy using a general-purpose AI, then paste the final version into Medium's editor manually. The MCP makes sense if the founder scales to a weekly cadence or brings on a co-author who needs independent access—then the OAuth and shared workspace start paying off.
Frequently asked
What does the Medium MCP do in Switchy?
The Medium MCP connects your team's Medium account to Switchy's AI workspace, letting you draft, publish, and manage Medium posts directly through AI conversations. You can ask the AI to create story drafts, update existing posts, or pull publication analytics without switching to Medium's web interface. It's useful for content teams who want AI assistance with their Medium publishing workflow.
Do I need to be a Medium publication owner to connect this?
You need a Medium account with OAuth permissions for the publication you want to manage. Medium's OAuth flow will ask you to authorise Switchy to act on your behalf — if you're connecting a publication (not just your personal account), you'll need editor or admin rights to that publication. Personal Medium accounts work fine for individual use cases.
Can the Medium MCP read my private drafts or unpublished stories?
Yes, if Medium's OAuth scopes permit it. The MCP can access whatever your Medium account can see, including drafts and unlisted posts. This is controlled by the permissions you grant during the OAuth flow. If you only want the AI to work with published content, you'll need to check Medium's API scope options before connecting.
Why use this instead of just writing in Medium's editor?
The MCP lets you use Switchy's AI to generate, edit, and iterate on Medium posts in a conversational workflow — useful if you're already using Switchy for research, drafting, or multi-step content projects. You lose Medium's native formatting tools and preview, so most teams use this for first drafts or bulk updates, then polish in Medium directly.
Does connecting Medium count against my Switchy workspace limits?
MCP connections don't count as separate seats, but API calls made through the Medium MCP consume your workspace's usage quota (if your plan has one). Heavy publishing workflows — like generating dozens of posts per day — will use more quota than occasional drafting. Check your plan's API limits if you're running high-volume content operations.