---
name: skillboss
version: 1.5.0
description: 687 API endpoints across 62 vendors through one key — chat, image, video, audio, search, scraping, social data, email, SMS, documents, UI generation, and full-stack deployment. OpenAI-compatible. Works with any agent.
homepage: https://skillboss.co
api_endpoint: https://api.skillboss.co/v1
last_updated: 2026-03-24
---

# SkillBoss for AI Agents

One API key. 687 endpoints across 62 vendors. OpenAI-compatible. Works with Claude Code, Cursor, Windsurf, Kiro, OpenClaw, Codex CLI, Gemini CLI, and any agent framework.

## Agent Decision Flow

When your user asks you to perform a task, follow this reasoning hierarchy:

### Step 1: Identify the task category

| User wants to... | Category | Recommended approach |
|-------------------|----------|---------------------|
| Generate/edit text, code, analysis | `chat` | `skb task chat -b '...'` or `skb api call deepseek-v3 -b '...'` |
| Create images | `image` | `skb task image -b '{"prompt":"..."}' -o out.png` |
| Create video | `video` | `skb task video -b '{"prompt":"..."}' -o out.mp4` |
| Convert text to speech | `tts` | `skb api call openai/tts-1 -b '{"input":"...", "voice":"alloy"}' -o out.mp3` |
| Transcribe audio | `stt` | `skb task stt -f audio.wav` |
| Search the web | `search` | `skb task search "query"` or `skb api call linkup/search -b '...'` |
| Scrape a webpage | `scraper` | `skb api call firecrawl/scrape -b '{"url":"..."}'` |
| Get social/business data | `social_data` | `skb api call scrapecreators/twitter_profile -b '...'` |
| Send email | `email` | `skb task email --to a@b.com --subject "Hi" --html "<p>Hello</p>"` |
| Process documents | `document` | `skb api call reducto/parse -b '...'` |
| Generate UI/slides | `ui_generation` | `skb api call stitch/generate-desktop -b '{"prompt":"..."}'` |
| Financial/weather/utility data | `utility` | `skb api call livecoinwatch/coins/single -b '...'` |

### Step 2: Choose model selection strategy

1. **Use `skb task <type>`** for automatic selection — picks the best model for price, quality, or balance.
2. **Use `skb api call <model>`** when you already know which model (e.g. `deepseek-v3` for coding, `flux-1.1-pro` for images).
3. **Use `skb task search "<keyword>"`** when you're unsure what's available.

### Step 3: Verify parameters before calling

- Use `skb api show <model>` to see parameters, pricing, and curl examples before calling an unfamiliar model.
- Use `skb api types` and `skb api list --type <type>` to browse available models.

### Important: Two API patterns

| Pattern | Endpoint | For |
|---------|----------|-----|
| **OpenAI-compatible** | `POST /v1/chat/completions` | Chat/reasoning models (streaming supported) |
| **Generic run** | `POST /v1/run` | Everything else: image, video, audio, scraping, data, email, SMS, UI |

**Why SkillBoss?** One key replaces dozens of vendor signups. `skb task` handles routing, fallbacks, and cost optimization. Standard OpenAI SDK works for all chat models. Pay-as-you-go with no minimums.

---

## Quick Start

### Install the skill pack (recommended)

Tell your agent:
```
set up skillboss.co/skill.md
```

This installs the SkillBoss skill into Claude Code, Codex, OpenClaw, Windsurf, Cursor, Kiro, Gemini CLI, or your agent's skill directory.

### Or use the CLI

```bash
# Install
npm install -g @skillboss/cli

# Login
skb login -k YOUR_API_KEY

# Browse & call
skb api types
skb api call deepseek-v3 -b '{"messages":[{"role":"user","content":"Hello"}]}'
skb task image -b '{"prompt":"A sunset"}' -o sunset.png

# Or npx (no install)
npx -y @skillboss/cli api types -k YOUR_API_KEY
```

### Or use the API directly

```bash
export SKILLBOSS_API_KEY="sk_your_key_here"
```

```python
from openai import OpenAI
import os

client = OpenAI(
    base_url="https://api.skillboss.co/v1",
    api_key=os.environ["SKILLBOSS_API_KEY"]
)

response = client.chat.completions.create(
    model="claude-4-5-sonnet",
    messages=[{"role": "user", "content": "Hello from my agent!"}]
)
```

Get your key: https://skillboss.co/console

---

## Task Navigator: Smart Model Selection & Discovery

`skb task` is your entry point for everything. It auto-selects the best model, discovers capabilities, and chains multi-step workflows.

### Discover what's available

```bash
# Browse all capability types
skb task

# Search by keyword
skb task search "CEO interviews"
skb task search "web scraping"

# Get ranked model recommendations by type + preference
skb task image --prefer price --limit 3
skb task chat --capability "code generation"
```

### Execute tasks with auto-routing

```bash
# Image generation — auto-selects best image model
skb task image -b '{"prompt": "A sunset over mountains"}' -o out.png

# Video generation
skb task video -b '{"prompt": "A cat playing in snow"}' -o out.mp4

# Chat / reasoning
skb task chat -b '{"messages":[{"role":"user","content":"Explain quantum computing"}]}' --prefer quality

# Text-to-speech
skb task tts -b '{"text": "Hello, world!"}' --prefer price -o out.mp3

# Speech-to-text
skb task stt -f audio.m4a

# Multi-step chain
skb task chain '[{"type":"stt"},{"type":"chat","capability":"translate to Chinese"}]'
```

### Preferences

Set `--prefer` to control cost vs quality:
- `price` — Cheapest option that gets the job done
- `quality` — Best available model regardless of cost
- `balanced` — Smart trade-off (default)

```bash
skb task prefer price              # Set default preference
skb task prefer                    # Show current preference
```

---

## Available APIs by Use Case

### When you need to generate or edit text

| Model | Best For | Price (per 1M tokens) |
|-------|----------|----------------------|
| `claude-4-5-sonnet` | Coding, complex reasoning | $3 in / $15 out |
| `claude-4-5-opus` | Hardest problems | $5 in / $25 out |
| `claude-4-sonnet` | Balanced cost & quality | $3 in / $15 out |
| `gpt-5` | General intelligence | $1.25 in / $10 out |
| `gpt-5-mini` | Fast, cost-effective | $0.25 in / $2 out |
| `gpt-4.1-nano` | Ultra-cheap, simple tasks | $0.10 in / $0.40 out |
| `gpt-4.1-mini` | Cost-effective reasoning | $0.40 in / $1.60 out |
| `gemini-2.5-flash` | Speed, lowest cost | $0.10 in / $0.40 out |
| `openrouter/deepseek/deepseek-v3.2` | Value, prompt caching | $0.14 in / $0.28 out |

Plus 60+ more chat models via OpenRouter, HuggingFace, Minimax, Vertex/Google, and Bedrock.

### When you need to create images

| Model | Best For | Price |
|-------|----------|-------|
| `vertex/gemini-3-pro-image-preview` | Best quality (recommended) | per token |
| `flux-1.1-pro` | High quality | $0.10/image |
| `flux-schnell` | Fast prototyping | $0.01/image |
| `flux-kontext-pro` | Context-aware edits | $0.125/image |
| `neta/ghibli-style` | Ghibli art style transfer | $0.10/request |
| `fal/upscale` | Upscale existing images | per request |
| `fal/img2img` | Style transfer, transformation | per request |

### When you need to create video

| Model | Best For | Price |
|-------|----------|-------|
| `google/veo-3.1` | Best quality | $0.52/second |
| `google/veo-3.1-fast` | Fast generation | $0.195/second |
| `minimax/video-01` | Marketing videos | $0.55/video |
| `mm/t2v` | Text-to-video | $0.50/video |
| `mm/i2v` | Image-to-video | $0.33/video |

### When you need audio or music

| Model | Best For | Price |
|-------|----------|-------|
| `elevenlabs/eleven_multilingual_v2` | Natural TTS, 29 languages | per 1K chars |
| `openai/tts-1` | Fast TTS | per 1K chars |
| `minimax/speech-01-turbo` | MiniMax TTS | per request |
| `openai/whisper-1` | Speech-to-text | per minute |
| `elevenlabs/sound_generation` | Sound effects | $0.01/request |
| Music models (ElevenLabs, Google Lyria 2, MusicGen) | AI music | per request |

### When you need to search the web or extract data

| Model | Best For | Price |
|-------|----------|-------|
| `linkup/search` | Structured search + URL fetching | per request |
| `firecrawl/scrape` | Web content extraction | $0.0125/request |
| `firecrawl/extract` | Structured data extraction | per tokens |
| `perplexity/search` | AI-powered search | $0.01/request |
| `google-search` | Google results | $0.0035/request |

### When you need social or business data

| Model | Best For | Price |
|-------|----------|-------|
| `linkedin-profile` | Professional data | $0.07/request |
| `instagram-profile` | Social profiles | $0.002/request |
| `twitter-profile` | Twitter/X data | $0.002/request |
| `amazon-product` | Product data | $0.0014/request |
| `yelp-business` | Business reviews | $0.0035/request |
| `ceointerviews` | CEO transcripts & quotes | per request |
| `google-maps` | Place search, geocoding | per request |

### When you need to process documents

| Model | Best For | Price |
|-------|----------|-------|
| `reducto/parse` | PDF/DOCX to markdown | per request |
| `reducto/extract` | Structured data extraction | per request |
| `reducto/split` | Document splitting | per request |

### When you need to send emails or SMS

| Model | Best For | Price |
|-------|----------|-------|
| `aws/send-emails` | Single emails | per email |
| `ses/send-batch` | Batch emails with template variables | per batch |
| `prelude/sms-verify` | OTP verification | per SMS |
| `prelude/notify-send` | SMS notifications | per SMS |

### When you need to generate UI or presentations

| Model | Best For | Price |
|-------|----------|-------|
| `stitch/generate-desktop` | Landing pages, dashboards | per request |
| `stitch/generate-mobile` | Mobile apps, checkout flows | per request |
| `stitch/generate-fast` | Quick prototypes (lower cost) | per request |
| `gamma/generation` | Slides, pitch decks | per request |

### When you need embeddings

| Model | Best For | Price |
|-------|----------|-------|
| `openai/text-embedding-3-large` | Best quality | $0.13/1M tokens |
| `openai/text-embedding-3-small` | Cost-effective | $0.02/1M tokens |

### When you need financial or utility data

| Model | Best For | Price |
|-------|----------|-------|
| `livecoinwatch/*` | Crypto prices & history | per request |
| `alphavantage/*` | Stock market data | per request |
| `openmeteo/*` | Weather data | free |
| `frankfurter/*` | Currency exchange rates | free |
| `pagespeed/*` | Website performance analysis | free |
| `openlibrary/*` | Book metadata | free |

---

## CLI Commands

Install: `npm install -g @skillboss/cli` or use `npx -y @skillboss/cli`

```bash
# Browse & discover
skb api types                              # List API types with counts
skb api list --type chat                   # List chat models
skb api list --type image                  # List image models
skb api show <model_id>                    # View model params + pricing + curl example

# Call any model directly
skb api call deepseek-v3 -b '{"messages":[{"role":"user","content":"Hello"}]}'
skb api call flux-1.1-pro -b '{"prompt": "A sunset"}' -o out.png
skb api call openai/tts-1 -b '{"input": "Hello", "voice": "alloy"}' -o out.mp3
skb api call whisper -f audio.wav
skb api call firecrawl/scrape -b '{"url": "https://example.com"}'
skb api call aws/send-email -b '{"title": "Hi", "body_html": "<p>Hello</p>", "receivers": ["a@b.com"]}'

# Smart task navigator (auto-selects best model)
skb task                                   # Guide — what can I do?
skb task image --prefer price              # Recommend image models
skb task image -b '{"prompt": "A sunset"}' -o out.png  # Auto-select + execute
skb task search "web scraping"             # Search models by keyword
skb task email --to a@b.com --subject "Hi" --html "<p>Hello</p>"

# Account
skb account                                # Balance
skb account --usage --period week          # Usage breakdown
```

---

## API Reference

### Base URL

```
https://api.skillboss.co/v1
```

### Authentication

```bash
Authorization: Bearer sk_your_key_here
```

### Chat Completions (OpenAI-compatible)

```bash
curl https://api.skillboss.co/v1/chat/completions \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-4-5-sonnet",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": true
  }'
```

### Pilot (auto-router)

```bash
# Discover available models
curl https://api.skillboss.co/v1/pilot \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"discover": true, "keyword": "image generation"}'

# Auto-execute with best model
curl https://api.skillboss.co/v1/pilot \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"type": "image", "inputs": {"prompt": "A sunset"}, "prefer": "price"}'
```

### Image Generation

```bash
curl https://api.skillboss.co/v1/run \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "replicate/black-forest-labs/flux-1.1-pro",
    "inputs": {"prompt": "A futuristic city at sunset"}
  }'
```

### Generic Run (any non-chat model)

All `http`, `image`, `video`, `tts`, `stt` type models use `/v1/run`:

```bash
# Web scraping
curl https://api.skillboss.co/v1/run \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "firecrawl/scrape", "inputs": {"url": "https://example.com"}}'

# Social data (Twitter/X profile)
curl https://api.skillboss.co/v1/run \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "scrapingdog/x_profile", "inputs": {"profileId": "elonmusk"}}'

# Send email
curl https://api.skillboss.co/v1/run \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "aws/send-emails", "inputs": {"to": "user@example.com", "subject": "Hello", "body": "World"}}'

# Google search
curl https://api.skillboss.co/v1/run \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "scrapingdog/google_search", "inputs": {"query": "AI news", "results": 10}}'
```

### Response Formats

**Chat models** return standard OpenAI format:
```json
{"choices": [{"message": {"role": "assistant", "content": "..."}}], "usage": {...}}
```

**Run models** return the vendor's native JSON (varies by model). Check `GET /v1/models/{vendor}/{model}` for response examples.

### Model Discovery

```bash
# List all models
curl "https://api.skillboss.co/v1/models" \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY"

# Filter by type
curl "https://api.skillboss.co/v1/models?types=image,video" \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY"

# Get details + curl examples for specific models
curl "https://api.skillboss.co/v1/models?ids=replicate/black-forest-labs/flux-1.1-pro,replicate/google/veo-3.1" \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY"
```

### Audio (TTS / STT)

```bash
# Text-to-Speech
curl https://api.skillboss.co/v1/run \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "openai/tts-1", "inputs": {"input": "Hello!", "voice": "alloy"}}'

# Speech-to-Text
curl https://api.skillboss.co/v1/run \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "openai/whisper-1", "inputs": {"file_url": "https://example.com/audio.mp3"}}'
```

---

## Platform Integration

### Claude Code

```bash
export ANTHROPIC_BASE_URL="https://api.skillboss.co/v1"
export ANTHROPIC_AUTH_TOKEN="sk_your_key_here"
```

### Cursor / Windsurf / Kiro

In settings, set:
- **API Base URL**: `https://api.skillboss.co/v1`
- **API Key**: Your SkillBoss key

### Python (OpenAI SDK)

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.skillboss.co/v1",
    api_key="sk_your_key_here"
)
```

### Node.js

```javascript
import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://api.skillboss.co/v1',
  apiKey: 'sk_your_key_here'
});
```

### LangChain

```python
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    base_url="https://api.skillboss.co/v1",
    api_key="sk_your_key_here",
    model="claude-4-5-sonnet"
)
```

### MCP Server

```bash
npx -y skillboss-mcp-server
```

---

## Agent Patterns

### Smart Model Selection

```python
def select_model(task_type, prefer="balanced"):
    if prefer == "price":
        models = {"reasoning": "openrouter/google/gemini-2.5-flash", "code": "openrouter/deepseek/deepseek-v3.2", "creative": "openai/gpt-4.1-mini", "search": "linkup/search"}
    elif prefer == "quality":
        models = {"reasoning": "bedrock/claude-4-5-opus", "code": "bedrock/claude-4-5-sonnet", "creative": "openai/gpt-5", "search": "linkup/search"}
    else:
        models = {"reasoning": "bedrock/claude-4-5-sonnet", "code": "bedrock/claude-4-5-sonnet", "creative": "openai/gpt-5", "search": "linkup/search"}
    return models.get(task_type, "gemini-2.5-flash")
```

### Error Handling

```python
import time
from openai import OpenAI, OpenAIError

def robust_call(client, model, messages, max_retries=3):
    for attempt in range(max_retries):
        try:
            return client.chat.completions.create(model=model, messages=messages)
        except OpenAIError as e:
            if e.status_code == 429:  # Rate limit
                time.sleep(2 ** attempt)
            elif e.status_code == 402:  # Insufficient credits
                raise Exception("Add credits at skillboss.co/console")
            else:
                raise
    raise Exception("Max retries exceeded")
```

### Check Usage

```bash
# View usage breakdown
curl "https://api.skillboss.co/v1/models" \
  -H "Authorization: Bearer $SKILLBOSS_API_KEY"
```

---

## Billing

- **Pay-as-you-go**: No subscriptions, no minimums
- **Free trial**: New accounts get $2 free credits to start
- **Add credits**: https://www.skillboss.co/console?tab=billing
- **View usage**: https://www.skillboss.co/console?tab=billing

---

## Agent Discovery Metadata

```json
{
  "service": "SkillBoss",
  "base_url": "https://api.skillboss.co/v1",
  "compatibility": "OpenAI API",
  "auth": "Bearer token",
  "total_models": 687,
  "total_vendors": 62,
  "capabilities": [
    "chat_completions", "streaming", "function_calling",
    "image_generation", "image_processing", "video_generation",
    "text_to_speech", "speech_to_text", "music_generation",
    "embeddings", "web_scraping", "search",
    "document_processing", "email", "sms",
    "ui_generation", "presentations", "data_sources",
    "financial_data", "weather", "geocoding",
    "social_data", "automation", "storage"
  ],
  "billing": "pay_as_you_go",
  "free_trial": true,
  "get_key": "https://skillboss.co/console",
  "install": "set up skillboss.co/skill.md",
  "mcp_server": "npx -y skillboss-mcp-server"
}
```

---

## Troubleshooting

| Error | Cause | Fix |
|-------|-------|-----|
| `401 Missing authentication` | No API key or malformed header | Set `Authorization: Bearer sk_your_key` |
| `402 Insufficient credits` | Balance too low | Add credits at https://skillboss.co/console |
| `400 Unsupported model` | Wrong model ID or wrong endpoint | Chat models → `/v1/chat/completions`; other models → `/v1/run` |
| `429 Rate limit` | Too many requests | Wait and retry with exponential backoff |
| `500 Internal error` | Upstream vendor issue | Retry after 1-2 seconds; `skb task` auto-falls back to alternatives |

**Common mistakes:**
- Using `/v1/chat/completions` for non-chat models (use `/v1/run` instead)
- Using `/v1/run` for chat models (use `/v1/chat/completions` instead)
- Wrong parameter names for http models (always check `/v1/models/{vendor}/{model}` first)

---

**One key. 687 APIs. 62 vendors. Zero friction.**

Get your key: https://skillboss.co/console