⛔ This repo has moved. The setup guide, SETUP.md, and the tutorial now live at bolloplayer/affinity-mcp-setup — point your AI agent there instead. This repo is kept only so old links don't 404.

Connect your AI to Affinity Photo

Affinity Photo v3 has a built-in MCP server. That means an AI agent — Claude Code, Codex, Antigravity, OpenCode — can write and run real scripts inside the app: adjustment layers, filters, batch work, all through Affinity's own JavaScript SDK. No extra software, no UI automation, no macOS requirement.

This repo exists so your agent can set that connection up for you. Point it here and it will find the config for whichever harness it's running in, plus the handful of non-obvious traps that otherwise cost an afternoon.

Just ask your AI

Open Affinity Photo, enable Edit ▸ Settings ▸ Model Context Protocol ▸ Enable Affinity MCP, then paste this into your agent:

Set up a connection to Affinity Photo's MCP server on Windows.

Follow github.com/bolloplayer/affinity-photo-claude-code-windows — read its SETUP.md and do what it says for the harness you're running in, including the verification at the end. If you get stuck, tell me where the doc stopped matching what you found.

That's it. Your agent reads SETUP.md, registers the connection, proves it works without touching anything of yours — then offers you a short menu: leave it there, run a colour-boost on your open photo and see the before/after, write a look of your own, or make the connection machine-wide.

⚠️ Expect one restart — it is not a failure

Partway through, your agent will stop and ask you to restart it. That is supposed to happen.

Every one of these tools loads its MCP configuration once, at startup. The agent has just written that configuration, so the connection cannot exist in the session that created it — there is no command that loads it live, in any of them.

Quit, start it again in the same folder, approve the affinity server if prompted, and say "continue". Your agent leaves itself a note, so the new session picks up where it stopped and goes straight to the before/after.

It is once for this folder. And once you've seen it working, your agent offers to register Affinity machine-wide, so you never do the setup again in any project.

Prefer to do it by hand? SETUP.md reads perfectly well as a human document too.

The connection, in one line

The endpoint is always:

http://[::1]:6767/sse

Harness	Config file	Transport
Claude Code — CLI, VS Code, Desktop Code tab	.mcp.json in the project	SSE native
Codex — CLI, ChatGPT Codex tab	~/.codex/config.toml	stdio bridge → SSE
Antigravity (agy) — Gemini	.agents/mcp_config.json (serverUrl)	SSE native
OpenCode — any model	opencode.jsonc or one CLI command	SSE native

Three things that look like bugs and aren't: it's [::1], never localhost or 127.0.0.1 (Affinity binds IPv6 loopback only); it's SSE only (there is no Streamable HTTP endpoint); and Affinity accepts MCP protocol 2025-11-25 alone, which is why Codex needs a bridge. SETUP.md has the full list and the reasoning.

What's here

SETUP.md	The instruction sheet — per-harness config, the rules not to "fix", the verification sequence, troubleshooting
examples/	A read-only connection check, and a two-layer Color Boost you can run on any photo
bridge/	The Codex protocol bridge, plus its smoke test
verify.ps1	Checks the plumbing before you start — Affinity running, port listening, handshake
CLAUDE.md	Connection internals, so Claude can diagnose problems from inside the chat
docs/sdk-notes.md	Confirmed SDK behaviours and mapped dead-ends

The tutorial

For a human walkthrough with screenshots — the Claude Code path end to end, and how to point Claude Code at DeepSeek instead without changing a single config file:

→ bolloplayer.github.io/affinity-photo-claude-code-windows

And if you haven't picked a model or a harness yet, docs/choosing-your-ai.md compares them — what's verified, what's assumed, and what each one costs.

Status

Verified on Windows 11 with Affinity Photo 3.2.x: Claude Code (CLI and VS Code), Codex CLI and the ChatGPT Codex tab, Antigravity, and OpenCode all connect, read the SDK preamble, and execute scripts. Affinity exposes the same server on macOS, but only Windows is verified here.

Affinity's AI Connector is a free beta and Canva-owned; its APIs and the scripting SDK may change without notice.

License

MIT — see LICENSE. Built against Affinity Photo's built-in MCP server (Affinity / Canva). Independent project, not affiliated with or endorsed by Canva or Affinity.