dora-mcp

Saga-Labs/dora-mcp
0 starsMITCommunity

Install to Claude Code

This server doesn't publish a one-line install command. Follow the setup in the source repository.

Summary

Multi-model AI image and video generator. 14 models behind one OAuth-secured MCP endpoint.

README.md

Dora MCP

Multi-model AI image and video generator, exposed as a hosted MCP connector for Claude.

A single OAuth-secured endpoint gives Claude access to 14 image and video models through one tool surface. Cheap-first defaults, async job polling, and full-size results stored on Firebase.

Models

| Kind | Model | Coin cost | Notes | |-------|--------------------|-----------|---------------------------------------------| | Image | z_image | 10 | Cheapest text-to-image | | Image | google_nano_banana | 40 | Low-cost text-to-image | | Image | google_nano_banana_edit | 40 | Low-cost editing from a reference photo | | Image | gpt4o_image | 60 | OpenAI GPT-4o image | | Image | gpt_image_2 | 60 | OpenAI gpt-image-2 | | Image | nano_banana_2 | 100–150 | Default — quality + reference photo | | Image | nano_banana_pro | 180–240 | Best face preservation | | Video | grok_text_to_video | 100–400 | Cheap text-to-video | | Video | grok_image_to_video | 100–400 | Cheap image-to-video | | Video | seedance_1_5_pro | 140–1800 | Default — solid quality, optional audio | | Video | kling_2_6 | 550–2200 | Image-to-video with audio | | Video | veo_3 | 600 | Premium text-to-video (8 s) | | Video | seedance_2_fast | 620–3960 | SeeDance 2 fast | | Video | seedance_2 | 760–12240 | SeeDance 2 — highest quality |

Tools

| Tool | Purpose | |------------------|-------------------------------------------------------------------------| | list_models | Returns all models cheap-first with per-call cost | | generate_image | Starts an image job; accepts prompt, model, resolution, aspect_ratio, reference_image_url | | generate_video | Starts a video job; accepts prompt, model, duration, resolution, aspect_ratio, with_audio, image_url | | check_job | Polls a job; returns inline image content for small assets, plus a Firebase Storage URL |

Async model: generate_* returns a job_id immediately. Call check_job until status is done. Image results are returned inline (base64) when the asset fits the API limit (~3.5 MB raw); larger images and all videos return a full-size URL.

Authentication

OAuth 2.1 with:

  • PKCE (RFC 7636)
  • Dynamic Client Registration (RFC 7591)
  • Refresh tokens

No pre-shared API keys. Standard OAuth flow handled by the MCP client.

Adding the connector

Claude Code

claude mcp add dora https://doravideo.com/mcp

You'll be prompted to authorize via browser the first time.

claude.ai (custom connector)

Settings → Connectors → Add custom connector → URL https://doravideo.com/mcp → authorize when prompted.

Pricing

Generations are paid in coins. Free tier gets a small starter balance; subscriptions add monthly coin packs. See doravideo.com/pricing.

Safety

Prompts are filtered for CSAM patterns at the API gateway before any model call. Generated assets are stored on Firebase Storage in the user's account scope.

Status

Live in production since May 2026. Same backend serves the Dora iOS app (paid product on the App Store) and the doravideo.com web app.

Contact

  • Email: sagalabs@proton.me
  • Issues: open an issue on this repo

License

MIT — see LICENSE.

Related MCP servers

Browse all →