How kRouter's Agent Skills auto-configure your AI tools

If you use Claude Code, Aider, or any terminal-based AI agent, you have probably wished it could do more than just edit files. What if your agent could also generate speech audio, run web searches, create embeddings for your RAG pipeline, or generate images -- all through the same local endpoint?

kRouter makes this possible with Agent Skills.

What Is a Skill?

A Skill is a markdown file hosted on GitHub that contains a structured description of one of kRouter's API endpoints. It includes the endpoint URL, authentication scheme, JSON request/response schema, supported models, error handling patterns, and working code examples.

When you feed this URL to an AI agent as a system prompt instruction, the agent reads the schema and instantly understands how to call that endpoint. No SDK installation. No manual API documentation. One URL and the agent is operational.

The 8 Skills

kRouter ships 8 skills out of the box, covering every endpoint category:

Detailed Skill Examples

krouter-chat: Chat and Code Generation

The most commonly used skill. Your agent sends a standard OpenAI chat completion request to localhost:20128 and kRouter routes it through the Zenith engine to the best available provider.

curl http://localhost:20128/v1/chat/completions \
  -H "Authorization: Bearer sk-krouter-local" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kr/claude-sonnet-4.5",
    "messages": [{"role": "user", "content": "Explain quicksort in Python"}]
  }'

krouter-tts: Text-to-Speech

Generate audio from text using any of the 4 supported TTS providers. See the full TTS routing guide for voice mapping details and cost comparison.

curl http://localhost:20128/v1/audio/speech \
  -H "Authorization: Bearer sk-krouter-local" \
  -d '{"model":"openai/tts-1","input":"Hello from kRouter","voice":"nova"}' \
  --output hello.mp3

krouter-stt: Speech-to-Text

Upload an audio file and get a transcript. kRouter translates between the OpenAI Whisper format and your configured STT providers.

curl http://localhost:20128/v1/audio/transcriptions \
  -H "Authorization: Bearer sk-krouter-local" \
  -F "[email protected]" \
  -F "model=whisper-1"

krouter-embeddings: Vector Embeddings

Generate embeddings for RAG pipelines, semantic search, or clustering. Same OpenAI SDK interface, routed to whatever embedding provider you have configured.

curl http://localhost:20128/v1/embeddings \
  -H "Authorization: Bearer sk-krouter-local" \
  -d '{"model":"text-embedding-3-small","input":"kRouter is an AI router"}'

krouter-web-search and krouter-web-fetch

These two skills give your agent access to the live web. web-search returns search results. web-fetch takes a URL and returns clean markdown -- perfect for feeding documentation pages into an agent's context.

# Web search
curl http://localhost:20128/v1/search \
  -H "Authorization: Bearer sk-krouter-local" \
  -d '{"query":"kRouter installation guide"}'
 
# Web fetch (URL to markdown)
curl http://localhost:20128/v1/web/fetch \
  -H "Authorization: Bearer sk-krouter-local" \
  -d '{"url":"https://docs.example.com/api"}'

How to Use Skills in Claude Code

Feed the raw GitHub URL of any skill to Claude Code as part of your prompt:

claude "Read this skill and use it to generate a podcast intro: \
  https://raw.githubusercontent.com/sifxprime/krouter/refs/heads/main/skills/krouter-tts/SKILL.md"

Claude reads the skill URL, understands the /v1/audio/speech endpoint structure, composes the JSON payload, calls localhost:20128, and saves the .mp3 to your working directory. No SDK installation. No API key management. It just works.

Chaining Skills in Multi-Step Workflows

The real power comes from chaining multiple skills. Here is a concrete agentic workflow:

1. Agent reads krouter-web-search skill -> searches for "latest React 19 breaking changes"
2. Agent reads krouter-web-fetch skill -> fetches the top 3 results as clean markdown
3. Agent reads krouter-chat skill -> summarizes the findings into a migration guide
4. Agent reads krouter-tts skill -> generates an audio briefing of the summary

Each step uses a different kRouter endpoint, but the agent needs zero custom integration code. It reads the skill files and calls the endpoints.

The MCP Server Mode

kRouter also runs as an MCP (Model Context Protocol) server, which means compatible agents can discover and call kRouter's endpoints automatically without even needing skill URLs:

krouter mcp

This starts kRouter in MCP stdio mode. Add it to your Claude Code MCP configuration, and every kRouter endpoint becomes a native tool that Claude can call directly. The agent does not need to read a skill file -- the MCP server advertises all available tools and their schemas at connection time.

Where to Find Skills

Open your local kRouter dashboard at http://localhost:20128/dashboard and click the Skills tab. Each skill shows a "Copy raw URL" button. That is the URL you feed to your agent.

Or browse them directly on GitHub: github.com/sifxprime/krouter/tree/main/skills.

Get Started

Install kRouter to unlock all 8 skill endpoints:

npm install -g @sifxprime/krouter

See the docs for endpoint configuration and the changelog for the latest skill additions.