Integration

GenMagic integration

GenMagic puts 450+ generative AI models from many labs (text, image, video, speech and music) behind one API key and one prepaid USD balance. Chat with any text model or write typed artifacts (writing, code, SVG, web page), generate or edit images and logos, make speech and music files, render video clips (create_video waits for the render), plan multi-asset campaigns for free with plan_assets, and browse the live model catalog with prices, daily catalog changes and benchmark leaderboards. Each generation draws its model's listed price from the balance (check it with get_usage); lookups are free. With brand personalization switched on in the GenMagic account, text, image, music and video outputs are steered on-brand (speech never is). Streaming chat and the MCP server are not exposed. Authentication is a GenMagic API key (gm_live_..., created on the API keys page of the account) sent as a Bearer token in the Authorization header.

Connect GenMagic How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in GenMagic.Built from the 15 GenMagic tools below, on the connection you add once.

What you can do with GenMagic

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Each badge shows the default agent permission of an endpoint (allowed: an agent calls it without asking; ask first: you are asked before an agent calls it, when the agent or chat asks permission for sensitive actions; never: no agent may call it), which every user can change endpoint by endpoint in their credential settings.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Benchmark, 2 operations

get_benchmarkGet benchmarkGETAgents: ask firstReturns one benchmark's definition and its model leaderboard, best first, each score with its source and a link to the original. Answers which model is best at a benchmark. Spends no credits.
list_benchmarksList benchmarksGETAgents: ask firstLists the benchmarks GenMagic tracks, each described the same way (what it tests, metric, status, maintainer, source); filter by category. Pass a benchmark id to get_benchmark for its leaderboard. Spends no credits.

Image, 2 operations

create_imageCreate imagePOSTAgents: ask firstGenerates 1 to 4 images from a prompt, or transforms a reference image given in image (https or data:image URL), and returns hosted URLs in data[].url. Billed per image at the model's listed price (list_models, category image).
edit_imageEdit imagePOSTAgents: ask firstEdits an uploaded image file following the instruction in prompt and returns the result in data[].url. Same engine, price and response as create_image with a reference image; masks are refused. Billed per image.

Model, 2 operations

get_modelGet modelGETAgents: ask firstReturns one model's full detail by id (for example openai/gpt-image-2): capabilities, pricing in its own unit (video: per-second price per resolution, clip lengths, sound) and its published benchmark scores. A lookup: spends no credits.
list_modelsList modelsGETAgents: ask firstLists GenMagic's live model catalog, newest first: each model's id (pass it as model), category, capabilities and pricing in its own unit. Filter with category, capability or search. A lookup: spends no credits.

Video, 2 operations

create_videoCreate videoPOSTAgents: ask firstRenders a video clip from a prompt and returns its url and cost_cents. Billed per second at the model's listed price, once, on completion; a failed render costs nothing. Waits for the job and returns the finished result; gives up after 10 minutes.
get_videoGet videoGETAgents: ask firstReads a video job by id (from create_video): status queued, processing, completed or failed, with the clip url and cost_cents once completed and error once failed. A lookup: spends no credits.

Asset plan, 1 operation

plan_assetsPlan assetsPOSTAgents: ask firstSends a conversation to GenMagic's studio agent, which replies in character and returns a plan of assets to make (modality, title, prompt). Not billed and generates nothing: run each prompt with the matching tool.

Chat completion, 1 operation

create_chat_completionCreate chat completionPOSTAgents: ask firstSends messages to a GenMagic text model (ids from list_models, category text) and returns the reply in choices[0].message.content, in OpenAI's format. Billed per token at the model's listed price.

Model change, 1 operation

list_model_changesList model changesGETAgents: ask firstLists what changed in the model catalog per UTC day over the last 1 to 90 days (default 7): new models, price changes with the price before and after, retirements, returns, changed details and benchmark updates. Spends no credits.

Music, 1 operation

create_musicCreate musicPOSTAgents: ask firstGenerates an original music track from a description (genre, mood, instruments, tempo) and returns it as a stored audio file. Billed at the music model's listed price (list_models, category audio).

Speech, 1 operation

create_speechCreate speechPOSTAgents: ask firstTurns text into spoken audio and returns it as a stored audio file in the model's own format (for example MP3). Billed at the speech model's listed price (list_models, category audio). The text is voiced exactly as given.

Text, 1 operation

create_textCreate textPOSTAgents: ask firstWrites a finished text artifact from a prompt and returns it in text: plain text, or set type to writing, code, svg (an SVG graphic) or website (a full web page). Billed per token at the model's listed price.

Usage, 1 operation

get_usageGet usageGETAgents: ask firstReturns the API key owner's live credit balance and lifetime spend (in US cents and USD) and the number of generations made. Spends no credits; check it before a large run to avoid a 402 refusal.

Use GenMagic from Claude Code, Cursor or Codex

Connect GenMagic in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its GenMagic operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect GenMagic

Create your workspace
Sign up free. Nothing to install.
Connect GenMagic once
Add your GenMagic connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.