Skip to main content
Glama

Clipia MCP

Create AI images, video, speech, music and presentations inside Claude, Cursor and ChatGPT — 60+ models in one MCP endpoint.

npm clipia-ai PyPI clipia License: MIT MCP

Clipia is an AI creation platform. This hosted, remote Model Context Protocol (MCP) server lets any MCP-capable agent generate images, video, speech and music; chat with text models; plan and compose videos; create and edit presentations; poll results; browse models; search prompt templates; and read your credit balance — no SDK and no code required.

Endpoint: https://mcp.clipia.ai/mcp — stateless Streamable HTTP, authenticated with a Clipia API key.


Quick start

Connect Clipia to Claude Code in one command (replace the placeholder with your key):

claude mcp add --transport http clipia https://mcp.clipia.ai/mcp \
  --header "Authorization: Bearer <CLIPIA_API_KEY>"

Then just ask in the chat: "Generate a neon city image with Clipia" — the preview lands right in the terminal.

Where to get a key: create one in the Clipia Developer Console. The key is shown once.

Sandbox without charges: keys with the clipia_test_ prefix run in a sandbox — instant mock results, no credits spent. Use a clipia_test_* key to validate your integration before going live.


Related MCP server: arcframe-mcp

Connect from any client

The endpoint is the same everywhere: https://mcp.clipia.ai/mcp. IDEs and CLIs authenticate with an API key from the Developer Console; claude.ai and ChatGPT connect by signing in to your Clipia account over OAuth (no key needed).

Ready-to-paste configs live in examples/.

Client

How to connect

Auth

Claude Code

claude mcp add --transport http clipia … (see Quick start)

API key

Claude Desktop

mcp-remote bridge in claude_desktop_config.json

API key

claude.ai (web/desktop/mobile)

Settings → Connectors → Add custom connector

OAuth

Cursor

~/.cursor/mcp.json (or project .cursor/mcp.json)

API key

VS Code

.vscode/mcp.json (servers + inputs)

API key

Cline

cline_mcp_settings.json (mcpServers)

API key

Windsurf

~/.codeium/windsurf/mcp_config.json (serverUrl)

API key

Codex CLI

~/.codex/config.toml (bearer_token_env_var)

API key (env)

Gemini CLI

gemini mcp add --transport http clipia …

API key

ChatGPT

Developer mode → Apps & Connectors → Create

OAuth

Claude Code

claude mcp add --transport http clipia https://mcp.clipia.ai/mcp \
  --header "Authorization: Bearer <CLIPIA_API_KEY>"

Claude Desktop

Claude Desktop bridges to remote HTTP servers through mcp-remote. Add to claude_desktop_config.json (Settings → Developer → Edit Config), then restart Claude Desktop:

{
  "mcpServers": {
    "clipia": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.clipia.ai/mcp",
        "--header",
        "Authorization: Bearer <CLIPIA_API_KEY>"
      ]
    }
  }
}

claude.ai (web / desktop / mobile)

  1. Open Settings → Connectors.

  2. Click Add custom connector and paste the URL: https://mcp.clipia.ai/mcp

  3. Click Connect and sign in to your Clipia account — no key needed (OAuth).

  4. In a new chat, ask for an image or a video: a live Clipia card with progress and the result appears inside the message.

Credits are charged to the connected Clipia account.

Cursor

Add to ~/.cursor/mcp.json (global) or .cursor/mcp.json (project), then restart Cursor:

{
  "mcpServers": {
    "clipia": {
      "url": "https://mcp.clipia.ai/mcp",
      "headers": { "Authorization": "Bearer <CLIPIA_API_KEY>" }
    }
  }
}

VS Code

Create .vscode/mcp.json in your workspace (or use the MCP: Add Server command). VS Code prompts for the key on first run and stores it securely:

{
  "servers": {
    "clipia": {
      "type": "http",
      "url": "https://mcp.clipia.ai/mcp",
      "headers": { "Authorization": "Bearer ${input:clipia_api_key}" }
    }
  },
  "inputs": [
    {
      "type": "promptString",
      "id": "clipia_api_key",
      "description": "Clipia API Key",
      "password": true
    }
  ]
}

Cline

Cline reads mcpServers from its cline_mcp_settings.json (MCP Servers → Configure MCP Servers in the Cline panel):

{
  "mcpServers": {
    "clipia": {
      "url": "https://mcp.clipia.ai/mcp",
      "headers": { "Authorization": "Bearer <CLIPIA_API_KEY>" }
    }
  }
}

Cline can also install this server itself — see llms-install.md.

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json (note the Windsurf-specific serverUrl field), then refresh Cascade → MCPs:

{
  "mcpServers": {
    "clipia": {
      "serverUrl": "https://mcp.clipia.ai/mcp",
      "headers": { "Authorization": "Bearer <CLIPIA_API_KEY>" }
    }
  }
}

Codex CLI

Add a block to ~/.codex/config.toml and export the key into your environment — Codex injects Authorization: Bearer from the variable automatically:

# ~/.codex/config.toml
[mcp_servers.clipia]
url = "https://mcp.clipia.ai/mcp"
bearer_token_env_var = "CLIPIA_API_KEY"

# in your environment:
# export CLIPIA_API_KEY=<CLIPIA_API_KEY>

Gemini CLI

gemini mcp add --transport http clipia https://mcp.clipia.ai/mcp \
  --header "Authorization: Bearer <CLIPIA_API_KEY>"

Verify the connection with gemini mcp list.

ChatGPT (Developer Mode)

  1. In ChatGPT (web): Settings → Apps & Connectors → Advanced settings → enable Developer mode (Plus, Pro, Business, Enterprise).

  2. Back in Apps & Connectors → Create. Name it Clipia, MCP Server URL: https://mcp.clipia.ai/mcp.

  3. Authentication — OAuth, tick "I trust this application", click Create.

  4. Sign in to your Clipia account in the popup. In a chat press + → More → Clipia, then ask in plain text.

Supported auth header schemes: Authorization: Bearer clipia_* (recommended), Authorization: Key <key>, or X-Api-Key: <key>.


Tools

The server always exposes 10 core tools to the agent. As of 2026-07-26, production exposes 15 agent tools: the 10 core tools plus chat, scenario planning, server-side video composition, presentation generation and presentation editing. Seven app-only helpers are hidden from the AI agent. Compact schemas keep the agent's context window light.

Tool

What it does

generate_image

Generate image(s) from a text prompt, optionally with reference images (editing / image-to-image). Waits briefly and usually returns the finished image inline (URL + small preview). Cost in credits is returned.

generate_video

Start a video generation from a text prompt (text-to-video) or from a start image (image-to-video, pass image_url). Returns request_id and cost in credits immediately — renders take 1–10 min, poll with wait_generation.

generate_audio

Generate speech from text with a selected voice and language. Returns an audio file when complete (mp3 or wav, depending on the model).

generate_music

Generate background music or a soundtrack from a description of mood, genre and tempo, or a song with your own lyrics.

wait_generation

Wait for a generation to finish (long-poll up to wait_seconds, then returns current status). Call repeatedly until COMPLETED, FAILED or CANCELED. Returns output URLs (and an inline preview) when done.

get_generation

Get the current status/result of a generation without waiting. When COMPLETED, output.images[].url is the inline webp preview and output.images[].original_url is the full-quality PNG/JPG.

list_models

List AI models with type (text/image/video/audio), capabilities and pricing in credits (per-1K rates for text models, base price for others). Filter by type / search; page with offset while truncated is true.

get_model

Get model details and pricing: input_schema for generation models; context limits and per-1K input/output rates for text models.

get_balance

Get the credit balance of the connected Clipia account and 30-day usage of the current API key.

search_templates

Search 3500+ curated prompt templates (hybrid text+semantic search, Russian or English query). Each result has a ready-to-use prompt and a recommended model.

chat

Currently enabled, feature-gated: chat with a text LLM using a prompt or messages array; returns reply text, token usage and credit cost.

generate_scenario

Currently enabled, feature-gated: turn a brief into per-scene video prompts and a soundtrack prompt.

compose_video

Currently enabled, feature-gated: stitch 2–20 completed scenes into a final MP4 with optional voiceover, soundtrack and subtitles.

generate_presentation

Currently enabled, feature-gated: render an editable PPTX, PDF and previews from a structured deck specification.

edit_presentation

Currently enabled, feature-gated: edit an existing deck, reorder or replace slides, change text and theme, and reuse unchanged illustrations.

app_get_generation

Internal / app-only: status poll used by the Clipia generation viewer card (MCP Apps). Hidden from the model; prefer get_generation.

Default models (used when no slug is passed): nano-banana-2 for images, seedance-2-fast-t2v / seedance-2-fast-i2v for video. Override with a model slug from list_models.

Parameters the model does not know. generate_image, generate_video, generate_audio and generate_music reject a call before the generation starts when input or a top-level argument carries a misspelled name, or a parameter the chosen model is known not to apply (for example seed on a model that does not declare it). The invalid_request error names the rejected parameters and lists the ones the model declares — check them with get_model and retry. Nothing is generated or reserved for a rejected call.


Why Clipia

  • Pay from Russia & CIS — Russian bank cards, SBP, MIR, no VPN required. Western processors reject these; Clipia is built for this market (and works internationally too).

  • 60+ flagship models, one endpoint — Western and Chinese models side by side: Veo 3.1, Kling 3, Seedance 2, Hailuo, Wan 2.7, Nano Banana, FLUX, Midjourney V7, Imagen 4, and more. No juggling multiple foreign subscriptions.

  • Sandbox without charges — clipia_test_* keys return instant mock results with no credit spend, perfect for wiring up an integration or CI.

  • Live preview in the chat (MCP Apps) — on claude.ai (web/desktop/mobile) every generation renders an interactive card with live progress, the finished media and an "Original" button. In Claude Code the preview lands inline in the terminal for vision-based iteration.

  • 3500+ prompt templates — search_templates gives the agent curated, ready-to-use prompts (hybrid search, RU/EN), each with a recommended model.

  • Cost in every response — every generation returns its exact cost in credits, and get_balance shows the remaining balance. No hidden MCP markup.

  • Compact tool surface — 10 core tools with tight schemas keep the agent's context window light; enabled capabilities add only the tools they need.


SDK

Prefer calling Clipia from your own code instead of an agent? Use the official SDKs against the same public API (async submit → status → result queue, credits-based billing):

  • TypeScript / Node: clipia-ai — npm install clipia-ai (also ships a clipia CLI).

  • Python: clipia — pip install clipia (sync + async clients).

Both support webhooks with HMAC-SHA256 signature verification and the same clipia_test_* sandbox keys.


Glama server runtime

The repository includes a minimal Docker runtime so Glama can start the public server entry and inspect its MCP tool definitions. The container is a stdio bridge to the hosted endpoint; it does not contain the Clipia application or any credentials.

Configure CLIPIA_API_KEY as a secret in Glama. Use a sandbox key with the clipia_test_ prefix so validation never spends credits. The optional CLIPIA_MCP_URL setting defaults to https://mcp.clipia.ai/mcp and only accepts HTTPS URLs.

docker build -t clipia-mcp-glama .
docker run --rm -i \
  -e CLIPIA_API_KEY='<CLIPIA_TEST_API_KEY>' \
  clipia-mcp-glama

Pricing

Billing is in credits from your connected Clipia account — the same balance as the website. Every generation returns its exact credit cost, and get_balance shows what's left. Different models cost different credits depending on resolution, duration and quality.

New accounts receive a small pack of welcome credits to evaluate the platform; after that a subscription is required. Four monthly plans (Basic / Standard / Pro / Ultima) — see clipia.ai/tariffs.

clipia_test_* sandbox keys return mock results with no credit charge — use them to develop and test without spending.


🇷🇺 Для России и СНГ

Clipia — AI-платформа для генерации изображений, видео, речи, музыки и презентаций, а также текстовых AI-чатов и сборки готовых роликов. Для пользователей из России и СНГ доступны оплата картой РФ, СБП и МИР, без VPN. 60+ западных и китайских моделей (Veo 3.1, Kling 3, Seedance 2, Nano Banana, FLUX, Midjourney V7) работают через один MCP-эндпоинт прямо из Claude, Cursor и ChatGPT.

Подключение в одну команду (ключ — в консоли разработчика):

claude mcp add --transport http clipia https://mcp.clipia.ai/mcp \
  --header "Authorization: Bearer <CLIPIA_API_KEY>"

Ключи clipia_test_* — песочница без списания кредитов для отладки. Биллинг в кредитах, стоимость каждой генерации возвращается в ответе. Тарифы: clipia.ai/tariffs.




License

MIT — see LICENSE. © 2026 Clipia (IP Zakharov Maksim Sergeevich).

Available Tools

15 tools
app_get_generationB
Read-onlyIdempotent
Inspect

Internal: status poll used by the Clipia generation viewer card. Prefer get_generation.

Inputs: request_id (required), include_preview.

chatAInspect

Chat with a text LLM (Claude / GPT / Gemini / DeepSeek class) and get the reply text plus token usage and credit cost. Pass either a single prompt or a full messages array. Charged in credits from the connected account.

Inputs: model, prompt, messages, max_tokens, +1 more.

compose_videoAInspect

Stitch 2–20 finished video scenes into ONE final clip server-side (scenes are normalized to a common frame/fps, concatenated, optional soundtrack replaces scene audio). Pass request_id of your COMPLETED video generations (or media.clipia.ai URLs). Returns request_id (cmp_*) — poll with wait_generation until COMPLETED, then output.video.url is the final mp4. Typical full-video flow: generate_scenario → generate_video per scene → compose_video.

Inputs: scenes (required), audio_url, normalize, subtitles, +5 more.

generate_audioAInspect

Generate speech from text (text-to-speech): pick a voice and language, get an mp3 back. Returns request_id and cost in credits immediately — narration is quick, poll with wait_generation until COMPLETED, then output.audio.url is the mp3. Pass the resulting request_id as voiceover_request_id in compose_video to narrate a multi-scene video, or use the mp3 on its own.

Inputs: text (required), input, model, voice, +4 more.

generate_imageAInspect

Generate image(s) from a text prompt, optionally with reference images (editing / image-to-image). Waits briefly for completion and usually returns the finished image inline (URL + small preview). Cost in credits is returned.

Inputs: prompt (required), seed, input, model, +8 more.

generate_musicAInspect

Generate background music / a soundtrack from a text description: instrumental (no vocals) by default — ideal to play under narration. Returns request_id and cost in credits immediately; music takes ~1–3 min, so poll with wait_generation until COMPLETED, then output.audio.url is the mp3. Pass the resulting request_id as audio_request_id in compose_video to score a multi-scene video, or use the mp3 on its own. Describe mood/genre/tempo/instruments in the prompt (English works best).

Inputs: prompt (required), input, model, instrumental, +2 more.

generate_presentationAInspect

Generate a slide presentation (editable PPTX + PDF + PNG previews) from a DeckSpec you compose. Returns request_id (prs_*) — poll with wait_generation until COMPLETED, then output has pptx_url, pdf_url, preview_urls[]. Write the full spec: title, language (ru/en), theme (clipia-dark default, clean-light for print/office), and 3–20 slides. Each slide has a "layout": cover, section, bullets, two-col, image-full, quote, stats, or closing. For illustrations set image.prompt (English) — generated by our image models and billed as credits; skip images for text-only slides. Costs credits (render fee + illustrations).

Inputs: title (required), slides (required), theme, language.

generate_scenarioAInspect

Plan a multi-scene video from a brief: an LLM director returns a structured scenario — per-scene English video prompts with durations, plus a soundtrack prompt. Feed each scene prompt to generate_video, then stitch the finished clips with compose_video. Charged in credits like chat (LLM usage).

Inputs: prompt (required), model, style, scene_count, +2 more.

generate_videoAInspect

Start a video generation from a text prompt (text-to-video) or from a start image (image-to-video, pass image_url). Returns request_id and cost in credits immediately — video renders take 1–10 minutes, poll with wait_generation.

Inputs: prompt (required), seed, input, model, +7 more.

get_balanceA
Read-onlyIdempotent
Inspect

Get the credit balance of the connected Clipia account and 30-day usage of the current API key.

Inputs: no named parameters.

get_generationA
Read-onlyIdempotent
Inspect

Get the current status/result of a generation without waiting. When COMPLETED, output.images[].url is the inline webp preview and output.images[].original_url is the full-quality PNG/JPG.

Inputs: request_id (required), include_preview.

get_modelA
Read-onlyIdempotent
Inspect

Get model details: supported input parameters (input_schema) and base price in credits.

Inputs: model (required).

list_modelsA
Read-onlyIdempotent
Inspect

List available AI models with type (image/video/audio), capabilities and base price in credits. Use before generating if the user asks for a specific model or to compare options.

Inputs: type, limit, search.

search_templatesA
Read-onlyIdempotent
Inspect

Search 3500+ curated prompt templates (hybrid text+semantic search, Russian or English query). Each result has a ready-to-use prompt and a recommended model — pass them to generate_image/generate_video.

Inputs: query (required), limit, media_type.

wait_generationA
Read-onlyIdempotent
Inspect

Wait for a generation to finish (long-poll up to wait_seconds, then returns current status). Call repeatedly until status is COMPLETED, FAILED or CANCELED. Returns output URLs (and an inline image preview) when done.

Inputs: request_id (required), wait_seconds, include_preview.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.9/5.0

    Scored across 15 tools

    Disambiguation4/5

    Most tools have clearly distinct purposes (e.g., generate_audio, generate_image, generate_video). However, app_get_generation and get_generation overlap significantly, and wait_generation adds a third polling tool, causing potential confusion.

    Naming Consistency4/5

    The majority follow a verb_noun pattern (generate_*, get_*, list_models, search_templates). The only inconsistency is app_get_generation, which uses an unnecessary prefix and deviates from the standard naming style.

    Tool Count5/5

    With 15 tools, the server covers a broad range of media generation tasks (audio, image, video, music, presentation, chat) and supporting utilities (balance, model info, templates). This count is well-scoped for its purpose.

    Completeness4/5

    Core generation and retrieval workflows are covered, but missing tools for listing or deleting generations and updating existing content (e.g., presentations) create minor gaps that agents may need to work around.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers