Pay any API - t2000

Agent Payments is x402 on Sui: the gateway at mpp.t2000.ai serves every major AI + data API as paid HTTP endpoints your agent pays in USDC. No signup, no API keys, no subscription — pay per call, gasless, settled on Sui in ~400ms.

npm install -g @t2000/cli
t2 init    # create a wallet
t2 fund    # transfer USDC to the printed address — $1 covers ~50 LLM calls

t2 pay https://mpp.t2000.ai/openai/v1/chat/completions \
  --data '{"model":"gpt-4o","messages":[{"role":"user","content":"Hello, world!"}]}'

Use t2 services to find an image-generation API, pay it with t2 pay to generate a hero image of a mountain lake at dawn, and show me the image URL and what it cost.

That’s a paid GPT-4o call for $0.02. The 402 challenge, payment, and retry are automatic. (Agent not set up yet? Use from your agent is one paste per client.) Browsing is free — no wallet, no account:

t2 services search "image"                         # find payable endpoints
t2 services inspect https://mpp.t2000.ai/openai    # exact prices + schemas

How it works

Client            Gateway (mpp.t2000.ai)         Upstream
  │  ── POST /openai/v1/chat/completions ──→  │
  │     (no Payment header)                    │
  │  ←── HTTP 402 + payment challenge ──      │
  │     (price, intent, recipient, expiry)    │
  │                                            │
  │  ── sign USDC tx + retry with Payment ─→  │
  │     header                                 │
  │                                       ──→  │  upstream call
  │                                       ←──  │  response
  │  ←── 200 + body ──                        │

USDC transfers go through 0x2::balance::send_funds — gasless via the Sui foundation sponsor. Your agent never needs SUI.

No charge on failure. Payment settles on-chain before the upstream runs, and if the upstream then fails, the gateway auto-refunds your USDC in the same request — net-zero, on-chain-verifiable. You only pay for calls that succeed.

What’s payable

LLMs, image generation, speech + transcription, search, scraping, weather, maps, crypto + stock data, translation, code execution, email, push — even physical mail. The live catalog is the source of truth:

Catalog	URL
JSON	`mpp.t2000.ai/api/services`
Agent-readable	`mpp.t2000.ai/llms.txt`
OpenAPI 3.1	`mpp.t2000.ai/openapi.json`
Web	`mpp.t2000.ai/services`

OpenAI-compatible inference

Every chat provider speaks the OpenAI Chat Completions shape — swap the path, keep the body. All $0.02/call:

Provider	Endpoint
OpenAI	`/openai/v1/chat/completions`
Groq	`/groq/v1/chat/completions`
DeepSeek	`/deepseek/v1/chat/completions`
Together	`/together/v1/chat/completions`
Mistral	`/mistral/v1/chat/completions`

embeddings, images/generations, audio/speech, and audio/transcriptions are mirrored too. The stock OpenAI SDK can’t sign a 402 — route the call through t2 pay or agent.pay() and parse the verbatim OpenAI response as usual.

Private inference, no key. t2000’s own Private Inference is a first-party x402 service — open + confidential (GPU-TEE) models, same pay-per-call flow.

Prices

Set per endpoint by each service; the live price is always in the catalog and the OpenAPI spec (x-payment-info.price). Ballpark: LLM calls

0.02–0.10 · images

0.02–0.40 · transcription

0.02–0.10 · search / scraping / code execution

0.02–0.20 · physical mail $2–3. Always pass --max-price <USD> (default $1.00) so a misconfigured endpoint can’t drain the wallet.

Three ways to pay

Surface	Call	Reference
CLI	`t2 pay <url> --data '…' --max-price 0.10`	CLI reference
MCP	`t2000_pay` + `t2000_services` from any AI client	Agent Wallet → MCP
SDK	`agent.pay({ url, body, maxPrice })`	Agent SDK

Lead a fresh session with “use t2 services.” In a brand-new chat, naming an external API directly (“generate an image via fal.ai”) makes some AI clients reply that they can’t reach it — they never consider the wallet. “Use t2 services to generate a hero image via fal.ai” loads the t2000_* tools from the first message; after one tool call the session routes correctly without the prefix.

Binary output (TTS audio, image bytes, PDFs) never streams back raw — the gateway hosts the artifact and returns { url, contentType, sizeBytes }, keeping bytes out of your agent’s context.

​How it works

​What’s payable

​OpenAI-compatible inference

​Prices

​Three ways to pay

​Links

How it works

What’s payable

OpenAI-compatible inference

Prices

Three ways to pay

Links