# PaleBlueDot.AI

> PaleBlueDot AI is a Palo Alto, California headquartered AI compute and intelligence platform
> supplying GPU infrastructure, dedicated clusters and model access at scale. Its developer-facing
> product is PBD TokenRouter, a unified AI model gateway that fronts 131 listed models from 22
> vendors behind a single API key and four vendor wire formats. This file was GENERATED by API
> Evangelist from the provider's public documentation and probed endpoints on 2026-08-26; the
> provider does not serve an llms.txt of its own.

## APIs

- [PBD TokenRouter Inference API](https://www.tokenrouter.com/docs/tokenrouter-feature-guide/): OpenAI-, Anthropic- and Gemini-compatible inference gateway. Base URL https://api.tokenrouter.com/v1 (use https://api.tokenrouter.com with no /v1 suffix for Anthropic Messages clients). Bearer auth with an sk- prefixed key.
- [PBD TokenRouter Management API](https://www.tokenrouter.com/docs/management-api-documentation/): Nine operations under https://api.tokenrouter.com/api/management for API key listing, enable/disable, spend quota, wallet balances and usage-log export. Bearer auth with a separate Management Key.

## Wire formats

- OpenAI Chat Completions: `POST /v1/chat/completions` (87 models)
- OpenAI Responses: `POST /v1/responses` (13 models)
- Anthropic Messages: `POST /v1/messages` (16 models)
- Google Gemini: `POST /v1beta/models/{model}:generateContent` (11 models)
- Image generation: `POST /v1/images/generations` (4 models)
- Video generation: `POST /v1/video/generations`, retrieve with `GET /video/generations/:task_id` (9 models)

## Machine-readable endpoints (no authentication required)

- [Model and price catalogue](https://api.tokenrouter.com/api/pricing): 132 models, 22 vendors, tiered pricing, video and image pricing, and the supported_endpoint path map. Served anonymously as JSON.
- [Platform configuration](https://api.tokenrouter.com/api/status): server address, quota-per-unit, currency display, enabled auth methods and console module flags. Served anonymously as JSON.

## Docs

- [Documentation home](https://www.tokenrouter.com/docs)
- [TokenRouter Feature Guide (getting started)](https://www.tokenrouter.com/docs/tokenrouter-feature-guide/)
- [Management API reference](https://www.tokenrouter.com/docs/management-api-documentation/)
- [FAQ](https://www.tokenrouter.com/docs/faq/)
- [Claude Code setup](https://www.tokenrouter.com/docs/claude-code-setup/)
- [Codex / ChatGPT setup](https://www.tokenrouter.com/docs/setup-tokenrouter-in-codex-chatgpt/)
- [OpenClaw setup](https://www.tokenrouter.com/docs/openclaw-setup/)
- [ZCode setup](https://www.tokenrouter.com/docs/zcode-setup/)
- [How to use image models](https://www.tokenrouter.com/docs/how-to-use-image-models/)
- [How to use video models](https://www.tokenrouter.com/docs/how-to-use-video-models/)
- [Model catalogue](https://www.tokenrouter.com/models/)
- [Release notes](https://www.tokenrouter.com/release-notes/)
- [Blog](https://www.tokenrouter.com/blog)

## Company

- [PaleBlueDot AI](https://www.palebluedot.ai/)
- [About and Security & Compliance](https://www.palebluedot.ai/about/)
- [Newsroom](https://www.palebluedot.ai/newsRoom)
- [Contact](https://www.tokenrouter.com/contact-us/) - support@tokenrouter.com
- [Discord community](https://discord.gg/XX57FY8S9X)
- [GitHub (TokenRouter-Tools)](https://github.com/TokenRouter-Tools)

## Legal

- [Terms of Use](https://www.tokenrouter.com/docs/terms-of-use/) - the TokenRouter platform is operated by Artemis Tokenrouter Inc.
- [Privacy Policy](https://www.tokenrouter.com/docs/privacy-policy/)
- [Global Data Processing Agreement](https://www.tokenrouter.com/docs/global-dpa/)
- [Conditions of Use](https://www.tokenrouter.com/docs/conditions-of-use/)

## Notes for agents

- There is NO published OpenAPI, AsyncAPI, GraphQL schema, MCP server or A2A agent card. Probed 2026-08-26; all missed.
- No `/.well-known/` document is served on any host. A status-code-only probe will see false 200s: three of the four hosts return catch-all bodies (Astro SPA shell, React SPA shell, or an empty 200).
- No rate-limit response headers are sent. Published limits exist only as a static docs table; exhaustion surfaces as a 429 with no Retry-After.
- Many Management API errors are returned with HTTP 200 and `success: false` in the body rather than a 4xx status. Do not branch on status code alone.
- No idempotency key mechanism is documented. Reversal is limited to the symmetric enable/disable key pair and re-setting a quota; no reversal window is stated anywhere.