Open LLMs plus image, music, voice, sound effect and video generation on one API key.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Bearer followed by your legit2compute API key (l2ck_...), from https://legit2compute.com/developers
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Your legit2compute account: credit balance, the media pipelines your key can run with their estimated cost, and per-model token prices. Call this first to see what is available.
This tool takes no arguments.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http mcp https://legit2compute.com/mcp \ --header 'Authorization: Bearer YOUR_TOKEN'
Headers
Authorization (required): Bearer followed by your legit2compute API key (l2ck_...), from https://legit2compute.com/developersYour legit2compute account: credit balance, the media pipelines your key can run with their estimated cost, and per-model token prices. Call this first to see what is available.
No arguments.
{
"type": "object",
"properties": {},
"additionalProperties": false
}Send a prompt (or a full OpenAI-style messages array) to a hosted language model and get the answer. Billed per token from your credit. Use model "l2c/auto" unless you need a specific one from list_models.
| Argument | Type | Description |
|---|---|---|
| model | string | Model id from list_models. Default "l2c/auto". |
| prompt | string | The user message. Ignored when messages is given. |
| system | string | Optional system prompt (with prompt). |
| messages | array | OpenAI chat messages: [{"role":"system|user|assistant","content":"..."}]. |
| max_tokens | integer | |
| temperature | number |
{
"type": "object",
"properties": {
"model": {
"type": "string",
"description": "Model id from list_models. Default \"l2c/auto\"."
},
"prompt": {
"type": "string",
"description": "The user message. Ignored when messages is given."
},
"system": {
"type": "string",
"description": "Optional system prompt (with prompt)."
},
"messages": {
"type": "array",
"items": {
"type": "object"
},
"description": "OpenAI chat messages: [{\"role\":\"system|user|assistant\",\"content\":\"...\"}]."
},
"max_tokens": {
"type": "integer",
"maximum": 32768,
"minimum": 1
},
"temperature": {
"type": "number",
"maximum": 2,
"minimum": 0
}
},
"additionalProperties": false
}Start a media job: images, image edits, music and songs, cloned voice and narration (in other languages too), sound effects, video from text or an image, long video to short clips, transcripts, articles, podcasts. Returns a job_id at once; then call get_job. Billed from your credit when it runs. Pipelines and example input: - image_generate: {"prompt":"a lighthouse at dawn, cinematic"} - image_edit: {"prompt":"make it night","input_image":"<https URL or a previous output key>"} - music_generate: {"prompt":"warm lo-fi piano","duration":30,"instrumental":true} - voice_generate: {"text":"Welcome back to the show.","voice":"maple:happy"} (optional "language":"es") - sound_effects: {"text":"ocean waves on a rocky shore","seconds":6} - video_t2v: {"prompt":"ocean waves at sunset","seconds":5} - video_i2v: {"prompt":"slow push-in","input_image":"<https URL or output key>"} - full_clip: {"url":"https://youtube.com/watch?v=...","num_clips":5} - transcribe_only: {"url":"https://...","language":"auto"} - article_gen: {"topic":"...","tone":"neutral"} - podcast_gen: {"topic":"...","minutes":10}
| Argument | Type | Description |
|---|---|---|
| input* | object | The pipeline input, e.g. {"prompt":"..."}. |
| title | string | Optional label for the job. |
| pipeline* | string | One of the pipelines listed above (account shows which your key can run). |
{
"type": "object",
"required": [
"pipeline",
"input"
],
"properties": {
"input": {
"type": "object",
"description": "The pipeline input, e.g. {\"prompt\":\"...\"}."
},
"title": {
"type": "string",
"description": "Optional label for the job."
},
"pipeline": {
"type": "string",
"description": "One of the pipelines listed above (account shows which your key can run)."
}
},
"additionalProperties": false
}Status of a media job. When it is completed, the result files with download links. wait_seconds (up to 50) keeps checking until the job ends or the time is up. Never submit a job again because it is still running: it is still billed.
| Argument | Type | Description |
|---|---|---|
| job_id* | string | |
| wait_seconds | integer | Wait for the job to finish, up to this long. Default 0. |
{
"type": "object",
"required": [
"job_id"
],
"properties": {
"job_id": {
"type": "string"
},
"wait_seconds": {
"type": "integer",
"maximum": 50,
"minimum": 0,
"description": "Wait for the job to finish, up to this long. Default 0."
}
},
"additionalProperties": false
}Every language model your key can call, with context length, tool support, modality (text or vision) and price. "l2c/auto" picks the cheapest model that fits each request.
No arguments.
{
"type": "object",
"properties": {},
"additionalProperties": false
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 22, 2026.