AI Tools Your Agent Can Call — Lightning-Native, No API Key

Most AI plumbing assumes a human is sitting at the other end. Email signups, OTP codes, credit-card pre-auths, monthly subscriptions, terms-of-service amendments. None of it makes sense when the user is an LLM agent that wants to do something — read a fax, place a phone call, narrate a book, transcribe an interview, translate a contract — without first having to prove it has a credit card.

Sats4AI is the tools layer for that agent. 41 AI capabilities, each callable in one HTTP request, paid per call in Lightning sats. No account, no API key, no signup. The payment is the credential.

This is the live catalog. It updates whenever we ship something new.

How to call#

MCP one-liner (Claude Code, Cursor, Cline, Continue, Mistral Le Chat, OpenClaw, anywhere MCP runs):

npx -y sats4ai-mcp

Direct L402 (any HTTP client, any language, any agent):

GET  https://sats4ai.com/.well-known/l402-services      # discovery manifest
POST https://sats4ai.com/api/l402/<endpoint>            # first call returns 402 + invoice

The first POST returns HTTP 402 with a Lightning invoice. Pay the invoice with any Lightning wallet. Resubmit the same request with the macaroon and preimage. Get the result. That is the entire dance — no out-of-band signup, no portal, no key rotation.

For agents specifically: pair this with CLW Cash or any Lightning-aware MCP wallet, and the agent pays for its own calls without a human in the loop.

What's in the catalog#

Communication#

Reach the world from a script. SMS, email, fax (send and receive), automated phone calls, and AI voice agents that can carry out a task on the line. The hands of the agent — when an LLM needs to talk to a human or a phone system.

• Send SMS — Send SMS messages worldwide. → /api/l402/send-sms · Dynamic (varies by destination country and message length)
• Place Phone Call — Place automated phone calls with text-to-speech or audio playback. → /api/l402/place-call · Dynamic (varies by destination and duration)
• AI Voice Agent Call — Send an AI voice agent to call any phone number and carry out a task on your behalf. → /api/l402/ai-call · From ~200 sats/min (dynamic, varies by duration and destination)
• Voice Bridge — Open a phone call that your own software drives one turn at a time. → /api/l402/voice-bridge-open · Deposit, per-minute by destination (unused time refunded)
• Send Email — Send a one-off email to any address. → /api/l402/send-email · 200 sats
• Send Fax — When your task requires a paper-trail on the other end (bank, notary, hotel in Japan, immigration office), send a fax to any number worldwide. → /api/l402/send-fax · 500 sats for ≤10 pages, +50 sats per additional page (max 350)
• Receive Fax — When expecting a fax back (bank confirmation, court filing, signed document), open a 24h receive window at our shared number +1 320 299 1523. → /api/l402/receive-fax · 500 sats base, 700 sats with OCR add-on

Speech & Audio#

Ears and voice. Transcribe any recording in 13 languages. Speak in 646 languages including 600+ underserved ones. Clone a voice from a sample. Narrate a full book into a chaptered audiobook. The mouth and ears of the agent.

• Text to Speech — Convert text to natural speech across 3 tiers. → /api/l402/text-to-speech · Per-character: OmniVoice ~106 chars/sat, Inworld Premium ~16 chars/sat, Minimax Studio ~5 chars/sat · OmniVoice Global (602+ langs), Inworld Premium (ELO 1217), Minimax Studio (voice cloning)
• AI Music Generation — Create original songs with AI vocals, or an instrumental by sending structure tags with no words. → /api/l402/generate-music · 5 sats/second (75-1500 sats for 15-300s) · Music-3
• Voice Cloning — Clone a voice you own or have permission to use, from an audio sample, for use with text-to-speech. → /api/l402/clone-voice · 7500 sats · MiniMax Voice Cloning
• Audio Transcription — Transcribe audio to text. → /api/l402/transcribe-audio · 10 sats/minute · Mistral Transcription
• Transcribe + Translate — Compound — audio in any of 13 source languages becomes both a transcript AND a translation in any of 119 target languages, in one payment. → /api/l402/transcribe-translate · Bundle: STT per minute + a translation leg priced by the target language (min 5 sats) · Mistral Voxtral + the engine the target language routes to
• Book to Audiobook — Convert EPUB, PDF, or TXT books into full AI-narrated audiobooks (M4B with chapters). → /api/l402/epub-audiobook · 500+ sats (dynamic, per-character) · OmniVoice Global (602+ langs), Inworld Premium (ELO 1217), Minimax Studio (voice cloning)

Image#

Generate, edit, analyze, restore, upscale, colorize, deblur, detect. Images in, images out. The eyes and the brush.

• AI Image Generation — Generate images from text prompts. → /api/l402/generate-image · 100-1200 sats (Standard 100, Better 150 at 1K / 300 at 2K, Top-ranked 1200) · xAI Grok Imagine 2, ByteDance Seedream 5 Pro, OpenAI GPT Image 2.5
• AI Image Editing — Edit images using AI — background removal, style transfer, object removal, and more. → /api/l402/edit-image · 200 sats (any output size) · MAI-Image-2.6
• AI Image Analysis — Analyze and describe image content using vision AI. → /api/l402/analyze-image · 21 sats · Qwen VL
• Background Removal — Remove image backgrounds with BiRefNet (SOTA). → /api/l402/remove-background · 44 sats · BiRefNet
• Image Upscaling — Upscale images 2-4x with AI super-resolution. → /api/l402/upscale-image · 5 sats · Real-ESRGAN
• Face Restoration — Restore and enhance faces in photos. → /api/l402/restore-face · 25 sats · CodeFormer
• NSFW Detection — Detect NSFW content in images. → /api/l402/detect-nsfw · 2 sats
• Object Detection — Detect and label objects in images with bounding boxes and confidence scores. → /api/l402/detect-objects · 5 sats · Grounding DINO
• Object Removal — Remove objects from images using AI inpainting. → /api/l402/remove-object · 130 sats
• Image Colorization — Colorize black-and-white or grayscale images with AI. → /api/l402/colorize-image · 5 sats
• Image Deblurring — Remove motion blur and defocus from images. → /api/l402/deblur-image · 110 sats · NAFNet
• Depth Estimation — Per-pixel depth map from a single photo. → /api/l402/estimate-depth · 5 sats · Depth Anything V2

Video#

Text-to-video and image-to-video, with optional audio. Async — submit, poll, retrieve.

• AI Video Generation — Generate video from text prompts. → /api/l402/generate-video · 250-400 sats/second by resolution (768p/2K) · MiniMax H3
• Image-to-Video Animation — Animate a still image into video with text guidance. → /api/l402/animate-image · 250-400 sats/second by resolution (768p/2K) · MiniMax H3

Text & Translation#

Frontier-class chat completion when an agent wants a tool-call to another model with no signup, and 119-language translation that auto-detects the source. Plumbing-grade.

• AI Text Generation — Generate text responses using frontier-class LLMs. → /api/l402/generate-text · per-character (BTC-pegged): Best ~10 chars/sat, Standard ~1000 chars/sat · Kimi K3 (Best), GPT-OSS-120B (Standard)
• Translate EPUB — A whole EPUB translated into another language, EPUB in and EPUB out. → /api/l402/translate-epub · Per visible character on the target language's engine rate (same as translate-text for that language), min 50 sats · The engine the target language routes to (see /api/languages)
• Text Translation — Translate text across 119 languages. → /api/l402/translate-text · 1+ sat (1000 chars/sat on the standard engine, min 1 sat; routed languages cost more — see /api/languages) · GPT-OSS 120B, or a higher-scoring model where one measurably beats it (the target language picks)
• Rare-Language Translation — Translate into 452 languages that frontier models do not serve - Bhojpuri, Maithili, Egyptian and Moroccan Arabic, Chhattisgarhi, Magahi, Manipuri, Kashmiri, plus indigenous and minority languages with no callable API anywhere. → /api/l402/translate-rare-language · 50 sats base + 0.002 sats/char
• Multilingual Ask — Ask a frontier AI a question in your own language and get the answer back in that same language, in one payment. → /api/l402/multilingual-ask · question length + 2,000-char answer allowance, min 2 sats

Documents & Files#

OCR for PDFs and images, structured receipt extraction, file conversion across 200+ formats, PDF merging, HTML→PDF rendering. Documents in, documents out.

• Document Extraction (OCR) — Extract text from PDFs and images with layout preservation. → /api/l402/extract-document · 10 sats/page · Mistral OCR
• Receipt & Invoice Extraction — Extract structured data from receipts and invoices. → /api/l402/extract-receipt · 25 sats/page · Mistral OCR + GPT-OSS-120B
• E-Signature — Send a document out for legally binding e-signature and have the signed copy emailed back. → /api/l402/e-signature · 1000 sats flat
• Boarding Pass Wallet — Turn airline boarding passes (PDF or screenshot) into Google Wallet passes with every field the barcode drops — boarding zone, gate, seat, boarding time. → /api/l402/boardingpass-wallet · 100 sats per pass
• File Conversion — Convert files between 200+ formats. → /api/l402/convert-file · 100 sats
• PDF Merge — Merge multiple PDF files into one document. → /api/l402/merge-pdfs · 100 sats
• HTML/Markdown to PDF — Convert HTML or Markdown content to a PDF document. → /api/l402/convert-html-to-pdf · 50 sats

3D#

Image-to-3D generation. GLB out.

• 3D Model Generation — Generate 3D models from images. → /api/l402/generate-3d-model · 1600 sats · Tencent Hunyuan3D

Other#

Misc capabilities.

• Render Card — Typographic card — headline, versus, list, stat or quote — rendered as PNG or JPEG. → /api/l402/render-card · 5 sats

Where to go next#

• Live machine-readable catalog: https://sats4ai.com/.well-known/l402-services
• MCP discovery: https://sats4ai.com/api/mcp/discovery
• Keyword search index (relevance-ranked): https://sats4ai.com/api/discover
• Why this exists: https://sats4ai.com/about
• MCP server source: https://github.com/cnghockey/sats4ai-mcp-server

---

Updated automatically from the L402 service catalog. Re-published as the catalog grows. Pubkey: 75a280ddc560c97edb5767720696bfce3c2677a3bc41574a8e77d4e59eec06a2.