Developer API
One API for every AI capability
OpenAI-compatible chat and reasoning, plus embeddings, reranking, OCR, speech-to-text, scraping, and live web search. North America-hosted, energy-measured, and private. One key for all of it.
Capabilities
Everything in one API
Every GreenPT capability is reachable through the same OpenAI-compatible API and a single key. Explore each below, or jump straight into the docs.
-
Chat Completions
OpenAI-compatible /v1/chat/completions for open-weight language and reasoning models.
(opens in a new tab) -
Streaming
Token-by-token responses over server-sent events for real-time chat and agents.
(opens in a new tab) -
Embeddings
green-embedding vectors for semantic search, clustering, and RAG pipelines.
(opens in a new tab) -
Reranker
green-rerank reorders retrieved results by true relevance, the last mile of search.
(opens in a new tab) -
Document OCR
Convert PDFs, Office files, and images into clean Markdown, JSON, or HTML.
-
Speech-to-Text
Transcribe pre-recorded files or live audio streams, multilingual and accurate.
-
Scraper
Extract clean, structured content from any web page in a single call.
-
Websearch
Run a live web search and get ranked results already scraped into AI-ready Markdown.
-
Green Coding
Code generation and completion models tuned for developer workflows.
(opens in a new tab) -
Models & Green Router
A catalog of open-weight models, with automatic routing to the most efficient one.
(opens in a new tab)
Why GreenPT
The measured AI API without the tradeoffs
Most AI APIs route your data through infrastructure you do not control and bill you for compute whose carbon cost stays hidden. GreenPT runs open-weight models on private, North America-hosted, energy-measured infrastructure, and never trains on your data.
-
OpenAI-compatible: switch the base URL and key, keep your existing SDKs.
-
One key for chat, embeddings, reranking, OCR, speech, scraping, and search.
-
Hosted in North America with the same GDPR principles and no-training policy.
-
Open-weight, self-hosted models: no lock-in to a closed provider.
-
The Green Router picks the most efficient model for each request.
-
Energy and carbon measured on every call.
The API, in short
Is the API OpenAI-compatible?
Yes. Chat completions follow the OpenAI /v1/chat/completions schema, so most existing SDKs and tools work by changing the base URL and key. Embeddings, reranking, OCR, speech-to-text, scraping, and web search are available through the same API.
What can I build with it?
Chat and reasoning, semantic search and RAG with embeddings and reranking, document OCR, speech-to-text, web scraping, live web search, and code generation. One key covers every capability.
Where is it hosted, and is my data private?
The API runs on private infrastructure in North America, with energy and carbon measured on every request. Models are open-weight and self-hosted. Your data stays in your region and is never used to train models.
Which models are available?
Open-weight language, reasoning, embedding, reranking, and speech models, served from a single catalog. The Green Router can pick the most efficient model for each request automatically.
See the model catalog →How do I get an API key?
Create an account to start a 14-day free trial, no credit card required, then generate a key from your dashboard and follow the Get started guide in the docs.
Read the API docs →
See the difference
Build on a measured AI API.
Start a free 14-day trial, no credit card. Get one key for chat, embeddings, reranking, OCR, speech-to-text, scraping, and web search, North America-hosted from the first request.
No credit card required.
- GDPR compliant
- CCPA-ready
- No training on your data