artokun / civitai

Civitai + comfyui-mcp

comfyui-mcp has native Civitai search built in (download_model action:"search_civitai")
plus the local half of the loop: download, wire, queue, tag. The full flow
(find, install, generate) needs no other server and no API key, and works on
every backend, including small local models behind the compact router.

The built-in flow (default path)

download_model({ action: "search_civitai", query, types: ["LORA"], base_models: ["Flux.1 D"] })
        │                    each hit: model_id · model_version_id · trigger words
        ▼
download_model({ action: "download_civitai", model_version_id, target_subfolder: "loras" })
        │
        ▼
list_local_models({ action: "list" })  →  panel_add_node loader / generate_image   # use it

• ALWAYS filter base_models when the user's checkpoint family is known.
  A Flux LoRA will not load on an SDXL checkpoint (see model-compatibility).
  CivitAI labels: Flux.1 D, SDXL 1.0, SD 1.5, Pony, Illustrious,
  Wan Video.

• target_subfolder must match the model type: checkpoints, loras, vae,
  controlnet, embeddings, upscale_models, and so on.

• Results are SFW-only by default (nsfw: true to opt in).
• Each hit carries trigger words. Put them in the prompt after installing.

• Prefer model_version_id over model_id. A Civitai page can list several
  versions and the user usually means a specific one.

Search needs no API key. CIVITAI_API_TOKEN (from
civitai.com/user/account) unlocks gated/early-access downloads and gated
search results. Set it once, in panel Settings under "Set CivitAI token…" or in env.

Recipes

"Find me a good anime LoRA for Flux and install it"

1. download_model({ action: "search_civitai", query: "anime style", types: ["LORA"], base_models: ["Flux.1 D"] }).

2. Present the top 3 to 5 with name, creator, base model, downloads, and the
   version id. Let the user pick.

3. download_model({ action: "download_civitai", model_version_id, target_subfolder: "loras" }).

4. In the panel, panel_add_node a LoraLoader, panel_set_widget the
   lora_name, wire it between checkpoint and sampler, and use the hit's
   trigger words in the prompt. Headless: generate_image(action="image") / build the workflow.

"Download this Civitai page for me" (user pastes a URL)

• Parse the modelVersionId from the URL if present; otherwise pass the model
  id from the URL to action:"download_civitai" (it resolves the latest version).
  A raw civitai.com/api/download/... URL also works via action:"download"
  with CIVITAI_API_TOKEN set.

Optional: the official Civitai MCP (community surface)

For features beyond search and install (browsing example images and their
generation params, collections, posting, reviews, bounties), pair the official
remote server. It is no longer auto-bundled; add it once:

claude mcp add --transport http civitai https://mcp.civitai.com/mcp \
  --header "Authorization: Bearer YOUR_CIVITAI_API_KEY"

Then mcp__civitai__* tools appear alongside comfyui-mcp's. Its discovery
results hand off identically (modelVersions[].id → action:"download_civitai").

That server also exposes write/social tools (post,
comment, review, DM, follow). Those publish on the user's behalf, so show
what you're about to post and get an explicit yes first. This skill covers
discovery, local install, and generation; don't post or message without being
asked.

See also

• model-registry for curated direct download URLs (HF + Civitai notes)
• model-compatibility for base-model / VAE / CLIP pairing (why a LoRA won't load)
• prompt-engineering because Civitai image params are a prompt goldmine

Sources

• Official: Civitai REST API https://civitai.com/api/v1 and Civitai MCP https://mcp.civitai.com/mcp
• Empirical: find→install→generate wiring is comfyui-mcp product guidance, not Civitai's prompting docs.