GenFire MCP for any AI
Connect GenFire to your AI assistant and generate images, videos, music, and full workflows directly from your prompts.
Open Claude settings
Launch the app or open claude.ai and go to
Settings → ConnectorsAdd the GenFire custom connector
Click Add custom connector, name it GenFire and paste the URL:
https://mcp.genfire.ai/mcpConnect and sign in
Click Add → Connect, sign in with your GenFire account — you're all set, now just ask Claude to generate an image.
The command set
37 tools, one connection. Every MCP client gets the same surface.
account
OAuth or API-key sign-in, credit balance, usage history, and live pricing.
catalog
Live catalog of 55+ models, published workflows, and voices. Inspect parameter schemas, defaults, and enums.
generate
The workhorse. Image, video, speech, music, and sound effects across every model.
edit
Post-production in a prompt — lipsync any video to audio, cut backgrounds, upscale, transcribe.
upload
Push a local image, video, or audio file into GenFire storage and plug it into any generation.
influencer
Trained, face-faithful characters you can @-mention in any prompt.
reels
Faceless reels end to end — script, voiceover, scenes, music, captions — plus recurring subscriptions that post on schedule.
runs · batches · workflows
Run multi-step pipelines, fan a prompt out across models, fetch and display any past output.
A complete creative studio inside your agent
Image generation, video creation, character training, and asset management in a single connection
> Generate a cinematic portrait of a woman standing alone in an empty subway
--model image.nano_banana_pro
--aspect 16:9
GenFire
Here she is. Dark coat draped over her shoulders, standing in an empty station.
→ portrait_01.png (4 variants)
Generative media
Image & video generation
From static stills to full cinematic video — your agent picks the model, sets the parameters, and generates. Nano Banana Pro, Seedream, and Recraft for images. Veo, Kling, Seedance, and Sora for video. Up to 4K, aspect ratios from 21:9 to 9:21.
> Score a moody synthwave track and voice this script, then lipsync it to the clip
GenFire
Generating audio…
✓ musicLyria 3 Pro · full song
✓ voiceoverElevenLabs v3
✓ lipsynclocked to clip.mp4
Audio engine
Music, speech & sound
Score it, voice it, sound-design it — without leaving the conversation. Lyria 3 Pro and ElevenLabs for music, multilingual speech in 30+ languages, cinematic sound effects, and lipsync that locks any video to any audio track.
> Make a 45-second true-crime reel and schedule one every morning
GenFire
Script written. Generating assets…
✓ voiceoverElevenLabs v3
✓ 8 scenesstyle-locked
✓ captionsburned in
✓ daily storyscheduled 9:00 AM
Ad engine
One prompt replaces your entire pipeline
Faceless reels end to end: AI script, voiceover, style-locked scenes, music, and burned-in captions in a vertical 9:16 short — then a recurring Story that posts a fresh one on schedule. From one brief.
> Generate @sarah walking through a Paris street market, golden hour
GenFire
Resolved @sarah → threading 4 reference photos. Switching to edit mode…
→ sarah_paris_01.png · identity locked
AI influencer
A face that never breaks
Train a face-faithful character from a few reference photos, then @-mention them in any prompt. Identity threads through reference images automatically — same face across every location, outfit, and model.
storyboardBrief, brand kit, script, shot list, storyboard
hook_pack3 monetizable UGC hooks plus body and CTA
ugc_adOutcome-oriented ad script, full spoken script
creative_packProduct-aware angles, hooks, monetizable script
product_ad_variantsMultiple ad script variants for one offer
Workflows
Pick a pipeline, skip the setup
Ready-made multi-step pipelines — storyboards, hook packs, UGC ad scripts, creative packs, ad variants. Your agent picks the right workflow and passes the inputs automatically. One ask, a coordinated set of outputs.
From idea to campaign in one conversation
Start small with a quick generation, or go deep with multi-model comparisons, character training, and full campaign workflows. Your agent handles the complexity.
Generate one asset in seconds
Describe what you need and your agent picks the best model, sets the parameters, and delivers the result. Images, videos, or both.
Generate a cinematic wide shot of a neon-lit Tokyo alley at nightBuild a visual system from a conversation
Train an influencer, generate scenes across locations and styles, produce videos, and manage everything from your generation history.
Train an influencer from these photos, then generate a 10-image lookbookCompare the best AI models side by side
Fan the same prompt out across Veo, Kling, Seedance, and Sora in one batch. Pick the winner and iterate from there.
Generate this scene on 4 different models and show me the resultsGot any questions left?
What is MCP?
The Model Context Protocol is an open standard that lets AI assistants call external tools. Connect the GenFire MCP server once and your assistant can generate images, videos, music, and full workflows mid-conversation — no copy-pasting between apps.
Do I need an API key?
Not always. Claude and ChatGPT sign in with OAuth — just paste the server URL and log in with your GenFire account. Cursor, Codex, and Gemini CLI use an API key from the developer dashboard, sent as a Bearer header.
What does it cost?
Generations through MCP draw from the same GenFire credits as the dashboard — identical pricing, no markup. The get_pricing tool returns live rates, so you can ask your agent what something costs before running it.
Which clients are supported?
Claude (web and desktop), Claude Code, ChatGPT, Cursor, Codex, Gemini CLI, Windsurf, VS Code Copilot — anything that speaks MCP over Streamable HTTP. One server, every client, all 37 tools.
MCP, CLI, or Skill — which should I use?
MCP is the universal option and the only one for chat apps like Claude and ChatGPT. In coding agents like Claude Code, Cursor, or Codex, the Skill drives the GenFire CLI directly with the same capability at roughly 95% lower per-call token cost.
Where do my generations go?
Every run lands in your GenFire library, same as generating from the dashboard. Fetch any past output later with get_run_output, list history with list_runs, or browse it all at genfire.ai.
Endpoint
URL
https://mcp.genfire.ai/mcpAuth
OAuth sign-in, or Authorization: Bearer YOUR_API_KEYNeed a key? Create one in the developer dashboard.