soundside.ai

MCP-native AI media generation with x402 pay-per-call. Image, video, audio, and music from 6 providers — composable via resource IDs. USDC on Base.

GitHub

Soundside — Developer Documentation

AI Media Production Platform for Agents

Soundside exposes 15 MCP tools for generating, editing, composing, extracting, and analyzing media — images, video, audio, music, text, and business artifacts. Connect any MCP client. Pay with an API key (credits) or crypto (x402 USDC on Base, no account needed).

Quick Start

# MCP endpoint
https://mcp.soundside.ai/mcp

# Auth: API key or x402 crypto payment
Authorization: Bearer <your-api-key>

POST https://mcp.soundside.ai/mcp
{"jsonrpc":"2.0","id":"1","method":"tools/list","params":{}}

Tools (15)

Generation

ToolWhat It DoesProviders
create_imageText-to-image, character referencesVertex AI, Grok, Runway, MiniMax, Luma
create_videoText-to-video, image-to-video, video extensionVertex AI (Veo 3.1), Runway, MiniMax, Luma, Grok
create_audioTTS, deprecated transcribe compatibility shim, voice cloning, sound effects, voice designMiniMax, Vertex AI, Runway, Creative Freedom
create_musicMusic from lyrics and style promptsMiniMax, Creative Freedom
create_textLLM chat completions, structured outputVertex AI (Gemini), Grok, MiniMax
create_artifactCharts, presentations, documents, diagrams; bundle mode for multi-artifact packages from a single briefPlotly, PPTX, WeasyPrint, Mermaid, Gamma

Editing & Analysis

ToolWhat It Does
edit_videoCore video transforms: trim, concat, crossfade, speed, loop, color grade, burn subtitles, custom FFmpeg
edit_audioMix, replace, or pad audio on existing media
compose_mediaAdd text, overlay media, or build split-screen composites
apply_effectKen Burns, speed ramp, film grain, vignette
extract_mediaExtract frames, frame sets, or audio tracks
analyze_mediaTechnical analysis, AI vision QA, transcription, segment detection, and EDL export

Library Management

ToolWhat It Does
lib_listBrowse projects, collections, resources, lineage, brand kits; query credit balance
lib_manageCRUD for projects, collections, resources, brand kits
lib_shareShare projects with other users by email

Pricing Philosophy

Soundside aims to break even on provider pass-through costs with a small margin (~10%). The real value is in the editing engine, library management, and self-hosted models — those are priced at $0.01/call ($0.03 for analyze_media vision_qa).

Live pricing is always available at:

GET https://mcp.soundside.ai/api/x402/status

This returns machine-readable per-tool, per-provider USDC prices. Prices are DB-driven and may change — always check the endpoint rather than hardcoding.

x402: Pay-Per-Call with Crypto

No API key needed. Pay with USDC on Base (L2) per tool call.

Network: eip155:8453 (Base mainnet)
Token: USDC
Facilitator: Coinbase CDP

See x402 Guide for full setup.

Guides

Examples

Links

Serveurs connexes