Canvora × Claude Cowork
On-brand visuals for Claude Cowork
Turn documents, plans and drafts into finished visual deliverables without leaving Claude — via Canvora's remote MCP server. OAuth sign-in, nothing to install.
Free credits, no card required. Please make sure there is always a human in the loop.
Connect in Claude
Settings -> Connectors -> Add custom connector
URL: https://api.canvora.ai/mcp
Sign in with your Canvora account (OAuth - no key handling).
Then, in any chat:
"Use Canvora to create a LinkedIn carousel about..."This is what your agent ships
One generation, localized twice. Same design, natively re-composed in Spanish and Arabic — right-to-left done correctly, not mirrored English.

Generated with the same commands shown on this page.
How it works
1. Describe the deliverable
Claude calls the chat tool to plan formats and style
You: "This product one-pager needs to become a pitch deck
and three LinkedIn posts. Brand: Dewstack."
Claude -> canvora.chat
proposes: pitch_deck (8 slides) + linkedin_post x3,
style: modern, brand kit: Dewstack2. Approve
Nothing generates until you confirm
You: "Go ahead."
Claude -> canvora.confirm_generation
Claude -> canvora.get_generation (polls status for you)3. Refine and export
Edits and exports stay in the conversation
You: "Slide 4 needs a bolder headline. Then export PPTX."
Claude -> canvora.edit_image (slide 4 revised)
Claude -> canvora.export_visuals (PPTX download link)What is Claude Cowork?
Claude Cowork is Anthropic's workspace where Claude handles multi-step knowledge work — documents, research, plans — with connectors to external tools. Canvora is the connector that turns that work into designed deliverables.
MCP-native
Claude connects to Canvora's remote MCP server and discovers all 16 tools on connect — generate, edit, localize, export — with full schemas.
OAuth, not keys
Sign in once in the connector dialog. No API keys to paste, nothing to host; access is revocable from your Canvora account at any time.
Confirm before spend
The chat-then-confirm flow means generation only starts when you approve the plan Claude proposes — the human stays in the loop by design.
A real Cowork session
From doc to deck in one thread
You: "Take the Q3 strategy doc and make the all-hands deck.
Our template style, 10 slides, end on the hiring CTA."
Claude: Using Canvora.
- chat: mapped the doc to 10 slides (agenda, 3 pillars,
metrics, roadmap, hiring CTA)
- awaiting your confirm... [you: "confirm"]
- confirm_generation -> generating
- get_generation -> completed, 10 slides
- "Slide 6's chart title is unclear" -> edit_image -> fixed
- export_visuals -> PPTX ready to download
One thread: source doc in, edited deck out.Localize the same deliverable
One deck, three markets, identical design
You: "Now the same deck in Spanish and Portuguese for the
regional teams."
Claude -> canvora.localize_generation (language: es)
Claude -> canvora.localize_generation (language: pt)
Same layout and brand; text natively re-composed per language.Why Claude Cowork + Canvora
- Remote MCP server — no installation, no hosting, OAuth sign-in
- 16 tools discovered on connect, schemas included: Claude knows what to do
- Explicit confirm step before any generation — human in the loop by design
- Edit, variations, localization and PPTX/PDF/ZIP export without leaving the chat
- Works on every Canvora plan, including Free
Command reference
Add --json to any command for machine-readable output. Progress goes to stderr, results to stdout.
| Command | Description |
|---|---|
| canvora generate | Create visuals from --idea, --input text, --url or --file-url; pick formats, brand, style, language |
| canvora status <id> | Check a generation; --json returns status and output file URLs |
| canvora download <id> | Save output files to a directory |
| canvora localize <id> | Same design, natively re-composed in another language (RTL included) |
| canvora variations <id> | Alternate takes on a finished generation |
| canvora edit <outputId> | Revise one visual with a natural-language instruction |
| canvora formats | List all 100+ format ids |
| canvora brands | List brand kits (colors, fonts, logo, tone) |
| canvora credits | Show credit balance |
100+ formats, one engine
Every format has platform-correct composition and safe zones — your agent picks by id.
Instagram Carousel
Up to 20 slides, hook to CTA
LinkedIn Carousel
Swipeable document posts
Pitch Deck
Coherent multi-slide sets
Instagram Post & Story
Feed and vertical formats
Stat Card
One number, designed to land
Infographic
Data and process visuals
Blog Header & OG Image
The launch surface
Ad Creatives
Display and social ads
FAQ
Do I need an API key? +
No. Claude connects over OAuth: add the connector URL, sign in with your Canvora account, done. Dynamic client registration means there is no manual onboarding step, and you can revoke access from your account at any time.
Can Claude spend credits without asking? +
No. Generation requires the explicit confirm_generation call, which only happens after you approve the plan in chat. Costs are flat per visual (10 credits) and per slide (15), so Claude can state the exact cost in the proposal.
Which tools does Claude get? +
All 16: chat, confirm_generation, get_generation, list_formats, list_brands, list_workspaces, edit_image, create_variation, localize_generation, convert_image, upload_photo, export_visuals, download_visual, get_edit_history, revert_edit, check_credits.
Does this work in Claude Code too? +
Yes — Claude Code can use the same MCP server, or go CLI-first: npx skills add canvora/canvora-mcp installs 12 skills and the same commands shown on our other agent pages.
Also works with
Your agent's next tool call is a design department.
Create a free account, set one key, and let Claude Cowork ship on-brand visuals.