Skills
Install the guid-ai-skills package to call multimodal generation from Agent conversations
Skills is Guid AI's multimodal generation skill package built on the V1 Open API. After installation, Agents can call image, video, and audio capabilities directly in conversation without writing HTTP requests yourself.
Current package name: guid-ai-skills
Supported Agents
| Platform | Notes |
|---|---|
| Claude | Claude Desktop, Claude Code, etc. |
| Cursor | Cursor IDE Agent |
| Codex | OpenAI Codex-style Agent environments |
| OpenClaw | Agents that support the Skills protocol |
| Hermes | Agents that support the Skills protocol |
| NemoClaw | Agents that support the Skills protocol |
Covered capabilities
| Category | Capabilities |
|---|---|
| Models | List available models by category |
| Images | Text-to-image |
| Videos | Standard video + guidai multimodal (7 modes) |
| Audio | Voice list, text-to-speech |
Does not include text chat (Chat / Completions / Responses).
Install & configure
1. Get an API Key
Create an API Key (sk-xxx) in the Guid AI admin console.
2. Install the skill package
- Obtain the
guid-ai-skills.ziparchive - Extract it into the Agent's Skills directory to get a
guid-ai-skills/folder - Ensure the directory contains
SKILL.md
Skills directory locations differ by platform. Common examples:
| Platform | Skills directory (example) |
|---|---|
| Cursor | Project- or user-level Skills directory (depends on client version) |
| Other Agents | See each platform's documentation for the Skills install path |
3. Configure environment variables
Fill in guid-ai-skills/.env, or export them as system environment variables:
GUIDAI_BASE_URL=https://www.guid.ai
GUIDAI_API_KEY=sk-xxx
| Variable | Description |
|---|---|
GUIDAI_BASE_URL | Gateway address without a path suffix |
GUIDAI_API_KEY | API Key |
4. Call from conversation
After configuration, describe your needs in conversation. The Agent follows the SKILL.md workflow to invoke the corresponding capability, for example:
- "Which image models are available?"
- "Generate a picture of a sunset beach"
- "Generate a video from this image"
- "Convert this text to speech"