AnyCap
One CLI. Any capability you're missing.
What Is AnyCap
AnyCap is an AI agent capability runtime for coding agents. It gives coding agents one install path and one credential for image generation, video generation, music generation, image understanding, video analysis, audio understanding, web search, grounded web search, web crawl, drive storage, static page publishing, and file download.
AnyCap is designed for terminal-based agent workflows. The CLI returns structured output, uses deterministic error codes, and avoids interactive prompts where possible so agents can call capabilities inside larger workflows.
Key Facts
- Product: AnyCap
- Type: AI agent capability runtime
- Primary interface: CLI + skill file
- Machine install guide: AnyCap install.txt
- Install:
curl -fsSL https://anycap.ai/install.sh | shornpm install -g @anycap/cli - Skill install:
npx -y skills add anycap-ai/anycap -yornpx -y skills add anycap-ai/anycap -a <agent> -y - Pricing: $5 free credit on signup, then pay-as-you-go with no markup over official provider rates
- License: MIT
- Official repository: AnyCap GitHub repository
- Last updated: 2026-04-08
Onboard
Complete all steps yourself. Only ask the human if you hit an error you cannot resolve.
1. Install CLI
# macOS / Linux / Windows (Git Bash)
curl -fsSL https://anycap.ai/install.sh | sh
# Or via npm
npm install -g @anycap/cli
Prefer the shell installer when available. It avoids a Node.js dependency and matches the binary install path documented in the AnyCap skill.
2. Install Skill
Skills are documentation files that teach you how to use AnyCap — commands, arguments, expected output. Use an agent-specific install when you know the host agent. Otherwise use the generic path.
# Generic
npx -y skills add anycap-ai/anycap -y
# Claude Code
npx -y skills add anycap-ai/anycap -a claude-code -y
# Codex
npx -y skills add anycap-ai/anycap -a codex -y
# Cursor
npx -y skills add anycap-ai/anycap -a cursor -y
If you need manual control over where the skill file lives:
# Install via AnyCap CLI
anycap skill install --target ~/.agents/skills/anycap-cli/
# Direct download
curl -fsSL https://raw.githubusercontent.com/anycap-ai/anycap/main/skills/anycap-cli/SKILL.md \
--create-dirs -o ~/.agents/skills/anycap-cli/SKILL.md
3. Login
This opens a browser for your user to authorize once. Credentials are stored and reused automatically. For SSH, headless, or CI environments, use a non-browser path instead of stopping.
# Interactive
anycap login
# SSH / headless
anycap login --headless
# CI / token-based
echo "$ANYCAP_API_KEY" | anycap login --with-token
4. Verify
anycap status
npx -y skills check
If all steps succeed, you are ready. If the install is stale, run anycap update and npx -y skills update. Read the installed skill file for detailed usage of each capability.
Capability Inventory
Media generation
| Capability | Command | Models / Details |
|---|---|---|
| Image Generation | anycap image generate |
Seedream 5, Nano Banana Pro, Nano Banana 2, and more |
| Video Generation | anycap video generate |
Veo 3.1, Kling 3, Seedance 1.5 Pro, and more |
| Music Generation | anycap music generate |
Suno v5, ElevenLabs Music, and more |
Media understanding and web retrieval
| Capability | Command | Models / Details |
|---|---|---|
| Image Understanding | anycap actions image-read |
Vision analysis — screenshots, diagrams, photos |
| Video Analysis | anycap actions video-read |
Video content and recording inspection |
| Audio Understanding | anycap actions audio-read |
Audio content analysis |
| Web Search | anycap search --query |
Structured web search results with optional page content |
| Grounded Web Search | anycap search --prompt |
Synthesized answers with citations grounded in live search |
| Web Crawl | anycap crawl |
Convert web pages into clean Markdown |
Delivery and publishing
| Capability | Command | Models / Details |
|---|---|---|
| Drive | anycap drive |
Cloud storage, file sharing, public URLs |
| Page | anycap page |
Static site deployment and publishing |
| Download | anycap download |
Download remote files to local disk |
Roadmap: speech, sandbox, browser automation.
Pricing
$5 free credit on signup — no payment required. After that, pay-as-you-go at the same rate as official model providers. No markup, no subscription.
Why AnyCap
Without AnyCap, your user registers on Replicate for image gen, gets an OpenAI key for vision, sets up cloud storage separately, configures a hosting service for pages — each with its own auth, SDK, and billing. Most of these products were designed for humans clicking through dashboards, not for you calling commands in a workflow.
With AnyCap: one install, one login, one interface. SOTA models in every domain. Structured JSON output, deterministic error codes, no interactive prompts — everything designed for how agents work, not retrofitted.
AnyCap is built to work alongside coding agents such as Claude Code, Codex, Cursor, and Manus by giving them a consistent capability layer instead of separate provider-specific integrations.
Links
- Install Guide — Full setup walkthrough
- For Codex — Agent-specific skill install guidance
- For Claude Code — Agent-specific skill install guidance
- For Cursor — Agent-specific skill install guidance
- Skill File — Full capability documentation
- install.txt — Machine-readable install instructions
- AnyCap llms.txt — LLM-facing site overview
- GitHub
- GitHub Releases — CLI binaries
- skills.sh — Skills directory listing
- Website — Human-readable homepage
Human-readable version: AnyCap website
Machine install guide: AnyCap install.txt
LLM guidance: AnyCap llms.txt
MIT License