Soundside

Connect your OpenClaw agent to Soundside's MCP tools for AI media generation, editing, library management, and technical analysis.

Security Scan

The skill's description and runtime instructions match an aggregator for media-generation tools, but the SKILL.md requires an API key and configuration that the registry metadata does not declare — an inconsistency you should resolve before installing.

Purpose & Capability

The skill claims to connect to Soundside's MCP for media generation and related tools — that matches the SKILL.md. However, the runtime instructions require a SOUNDSIDE_API_KEY and a specific mcpServers URL, while the registry metadata declares no required environment variables or primary credential. The missing declaration is an incoherence (the skill needs a credential but the metadata doesn't list it).

Instruction Scope

The SKILL.md is instruction-only and scoped to configuring openclaw.json to point at https://mcp.soundside.ai/mcp with an Authorization: Bearer ${SOUNDSIDE_API_KEY} header and restarting the gateway. It does not instruct reading unrelated files or exfiltrating other environment variables, so the behavior described stays within the stated purpose. It does instruct you to store the API key in openclaw.json (a plaintext config file), which has security implications to consider.

Install Mechanism

There is no install spec and no code files. This is lower-risk from an installation perspective because nothing is downloaded or written by the skill itself.

Credentials

The SKILL.md expects a SOUNDSIDE_API_KEY to be stored in openclaw.json and used as a Bearer token, but the registry metadata lists no required env vars or primary credential. The single credential requested is proportional to the described capability, but it should be declared in the metadata. Also note the SKILL.md references many downstream providers (Vertex AI, Runway, etc.); it's unclear whether Soundside requires additional provider credentials or will proxy calls (and whether your data will be forwarded to third parties).

Persistence & Privilege

The skill is not marked always:true and is user-invocable; it does not request persistent platform privileges beyond standard agent skill behavior.

What to consider before installing

Before installing: 1) Verify the publisher and domain (mcp.soundside.ai) and confirm the GitHub docs links are legitimate and match the service. 2) Confirm that you need to provide SOUNDSIDE_API_KEY (the registry metadata should declare this); ask the publisher to update the metadata if it is missing. 3) Consider how and where you store the API key — openclaw.json appears to hold it in plaintext; use a secrets mechanism if available. 4) Ask whether Soundside will proxy calls to third-party providers (Vertex, Runway, etc.) and whether those providers or Soundside may see or retain your media or metadata; verify pricing and billing behavior to avoid unexpected charges. 5) If you cannot validate the publisher or the docs, avoid adding the API key or enabling the skill until you receive clearer metadata and an explicit trust signal from the vendor.

Soundside — OpenClaw Skill

Connect your OpenClaw agent to Soundside's 12 MCP tools for AI media generation, editing, and analysis.

Setup

Add to your openclaw.json:

{
  "skills": {
    "entries": {
      "soundside": {
        "enabled": true,
        "env": {
          "SOUNDSIDE_API_KEY": "mcp_your_key_here"
        }
      }
    }
  },
  "mcpServers": {
    "soundside": {
      "transport": "streamable-http",
      "url": "https://mcp.soundside.ai/mcp",
      "headers": {
        "Authorization": "Bearer ${SOUNDSIDE_API_KEY}"
      }
    }
  }
}

Then restart: openclaw gateway restart

What You Get

Once connected, your agent has access to:

Generation (6 tools)

• create_image — Text-to-image across 5 providers (Vertex AI, Grok, Runway, MiniMax, Luma)
• create_video — Text/image-to-video across 5 providers (Vertex Veo 3.1, Runway, MiniMax, Luma, Grok)
• create_audio — TTS, transcription, voice cloning (MiniMax, Vertex AI)
• create_music — Music from lyrics + style prompt (MiniMax)
• create_text — LLM completions with structured output (Vertex Gemini, Grok, MiniMax)
• create_artifact — Charts, presentations, documents, diagrams

Editing & Analysis (2 tools)

• edit_video — 21 editing actions: trim, concat, Ken Burns, mix audio, text overlays, color grading, film grain, split screen, and more
• analyze_media — Technical analysis + AI vision QA scoring

Library (3 tools)

• lib_list — Browse projects, collections, resources
• lib_manage — Create/update/delete library entities
• lib_share — Share projects by email

Durable Resource Pattern

Every generation returns a resource_id that persists across sessions:

1. Generate media → receive resource_id
2. Chain into editing: edit_video(resource_id=..., action="add_text", ...)
3. Organize: lib_manage(entity_type="project", operation="create", ...)
4. Check status: lib_list(entity_type="resources", resource_id=...)
5. Download only for final delivery

This keeps workflow state durable without local storage.

Example Workflows

Generate and edit an image:

"Create an image of a sunset over the ocean using Vertex AI,
 then add the text 'Golden Hour' as a title overlay"

Produce a narrated video:

"Generate a video of waves crashing using Luma,
 create TTS narration saying 'The ocean calls to those who listen',
 then mix the narration into the video"

Build a presentation:

"Create a pitch deck with 5 slides covering our Q1 metrics,
 include a revenue chart showing growth from $100K to $600K"

Pricing

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

Soundside charges near-cost on provider pass-through (~10% margin). Editing and analysis are $0.01/call. A typical video pipeline (image → video → edit → analyze) costs $0.50-3.00 depending on provider.

Docs

• Getting Started
• Tool Reference
• x402 Pay-Per-Call