# ODEN > ODEN builds EU-native retrieval infrastructure. Two products: the ODEN Search > API, which turns one HTTP call into a citation-backed JSON answer for LLMs and > AI agents, and Tungsten, a local-first AI code editor that runs its model on > the user's own machine. ODEN runs on Cloudflare Workers and is operated from Sweden. The Search API is a direct alternative to Tavily and Firecrawl; robots.txt and the EU TDM opt-out are honored at the source, which is the part most retrieval APIs skip. Full product description in one file: [https://oden-api.com/llms-full.txt](https://oden-api.com/llms-full.txt) Crawling and training signals: [https://oden-api.com/ai.txt](https://oden-api.com/ai.txt) ## Products - [ODEN — AI search API for LLMs](https://oden-api.com/): What ODEN is, what the API returns, and how to get a key. - [ODEN AI — Swarm Search & AI Research Assistant](https://oden-api.com/oden-ai): Conversational web research, multi-model reasoning, and citation-backed synthesis. - [Tungsten — local, private AI code editor](https://oden-api.com/tungsten): A VS Code fork with a local model in the box. Free, Windows, no telemetry. ## The Search API - [Search endpoint](https://api.oden-api.com/search): POST a query, receive `answer` plus ranked `citations` as JSON. Auth: `Authorization: Bearer `. - [Health](https://api.oden-api.com/health): service status, no auth. - [ODEN API documentation](https://oden-api.com/docs): Request and response shape, parameters, auth, error codes, examples. - [Get an API key](https://oden-api.com/app): sign in and generate a key. Free tier: 1,000 searches/month. ## Tungsten, the editor - [Product page](https://oden-api.com/tungsten): what it is, what runs locally, system requirements. - [Download for Windows](https://github.com/Samiestenberg/tungsten/releases/latest): free, MIT licensed. - [Source on GitHub](https://github.com/Samiestenberg/tungsten): fork of Visual Studio Code. ## Pricing - [ODEN pricing](https://oden-api.com/pricing): Free tier, Pro at EUR 29/month, and one-time top-ups that never expire. - Free: 1,000 searches/month, no card. - Pro: EUR 29/month for 6,000 searches/month. - Top-up: EUR 7 per additional 1,000 searches, one-time, never expires. - Tungsten: free, and stays free — it costs us nothing to run because it runs on your machine. ## Documentation - [Quickstart](https://oden-api.com/docs/quickstart): Make your first ODEN API call in five minutes. Get a key, send a query, and read the synthesized answer and citations — in curl, Python, JavaScript and TypeScript. - [Authentication](https://oden-api.com/docs/authentication): How to authenticate with the ODEN API: bearer-token keys in the Authorization header, where to create and rotate keys, and how to keep them safe. - [Search endpoint](https://oden-api.com/docs/search): Full reference for POST /search: the request parameters (query, answer, depth, include_snippets) and every field in the JSON response, with examples. - [Answers and citations](https://oden-api.com/docs/answers-and-citations): How ODEN returns synthesized answers and ranked citations, when to request each, and how to attribute sources in your LLM responses correctly. - [Rate limits and quotas](https://oden-api.com/docs/rate-limits): How ODEN rate limits requests per key, what a 429 means, how quotas map to Free, Pro and top-ups, and how to back off and retry cleanly. - [Errors](https://oden-api.com/docs/errors): Every ODEN API error code and how to handle it: 400 invalid body, 401 unauthorized, 429 rate limit, and 5xx. Includes the error response shape. - [SDKs and clients](https://oden-api.com/docs/sdks): ODEN is a plain JSON-over-HTTPS API, so no SDK is required. Copy-paste a minimal Python or TypeScript client, or call it directly from any language. - [Changelog](https://oden-api.com/docs/changelog): Version history for the ODEN web search API — endpoints, parameters and response changes, newest first, so you can track what changed and when. ## Comparisons - [ODEN vs Tavily — 2026 comparison for LLM developers](https://oden-api.com/vs/tavily): ODEN vs Tavily in 2026: both return a synthesized answer plus citations for LLMs. How they compare on pricing, EU hosting, TDM compliance and latency. - [ODEN vs SerpAPI — 2026 comparison for LLM developers](https://oden-api.com/vs/serpapi): ODEN vs SerpAPI in 2026: SerpAPI returns raw search-engine results; ODEN returns a synthesized answer with citations. Which fits an LLM pipeline, and the cost. - [ODEN vs Serper — 2026 comparison for LLM developers](https://oden-api.com/vs/serper): ODEN vs Serper in 2026: Serper is a fast, cheap Google SERP API; ODEN synthesizes an answer with citations. Cost, output shape and which suits RAG. - [ODEN vs Bing Search API — the 2026 replacement](https://oden-api.com/vs/bing-search-api): Microsoft retired the Bing Search API in August 2025. ODEN is a drop-in replacement for LLMs: a synthesized answer plus citations from one EU-hosted call. - [ODEN vs Brave Search API — 2026 comparison](https://oden-api.com/vs/brave-search-api): ODEN vs Brave Search API in 2026: Brave returns results from its independent index; ODEN synthesizes an answer with citations for LLMs. Hosting, cost and fit. - [ODEN vs Google Custom Search — 2026 comparison](https://oden-api.com/vs/google-cse): ODEN vs Google Custom Search (CSE / Programmable Search) in 2026: CSE returns capped, custom-scoped results; ODEN synthesizes an answer with citations for LLMs. ## Use cases - [Build a RAG pipeline with real-time web search](https://oden-api.com/use-cases/rag): Add real-time web retrieval to a RAG pipeline with ODEN: one call returns context-ready text and citations, so your model answers from the live web. - [Add ODEN web search to LangChain](https://oden-api.com/use-cases/langchain): Wire ODEN into LangChain as a tool or retriever in a few lines: give your agent real-time web search that returns a synthesized answer plus citations. - [OpenAI function calling with web search](https://oden-api.com/use-cases/openai-function-calling): Give an OpenAI model real-time web search with ODEN via function calling: define one tool, run the call, and return a synthesized answer with citations. - [Claude tool use with web search](https://oden-api.com/use-cases/claude-tool-use): Give Claude real-time web search with ODEN through tool use: define one tool, handle the tool_use block, and return a synthesized answer with citations. - [LlamaIndex web search integration](https://oden-api.com/use-cases/llamaindex): Add ODEN web search to LlamaIndex as a tool or custom retriever: bring real-time web knowledge, with citations, into your LlamaIndex agents and query engines. - [Add citations to your LLM responses](https://oden-api.com/use-cases/citations): Make your LLM cite its sources with ODEN: every answer comes with ranked, attributable citations, so you can show users where each claim came from. - [GDPR-compliant web search API for the EU AI Act](https://oden-api.com/use-cases/eu-compliance): ODEN is an EU-hosted web search API that honors robots.txt and the DSM Article 4 TDM opt-out, helping your AI application meet GDPR and EU AI Act expectations. ## Writing - [ODEN blog](https://oden-api.com/blog): Engineering notes on retrieval, EU compliance, and local AI tooling. - [ODEN Search: The Sub-Second Web Retrieval & Indexing Engine](https://oden-api.com/blog/oden-search-sub-second-web-retrieval-engine-for-llms): ODEN Search is a high-speed real-time web retrieval engine delivering clean citations and sub-second index queries for AI models and modern web apps. (2026-08-04) - [Introducing ODEN AI: Next-Gen Reasoning, Live Web Grounding & Privacy](https://oden-api.com/blog/introducing-oden-ai-real-time-reasoning-and-web-grounding): ODEN AI combines deep reasoning models, instant web retrieval, multi-modal context, and multi-tenant privacy. Experience real-time intelligence without limits. (2026-08-04) - [Introducing Tungsten: a local, private AI code editor](https://oden-api.com/blog/introducing-tungsten-a-local-private-ai-code-editor): Tungsten is a free, privacy-first fork of VS Code with two local AI models inside the app — no cloud, no telemetry, no account. Here's what it does and why. (2026-07-30) ## Company and policy - [Contact ODEN](https://oden-api.com/contact): Support and sales contact. - [OdenBot — our crawler](https://oden-api.com/bot): What OdenBot is, how it identifies itself, and how to block it. - [Privacy policy](https://oden-api.com/privacy): What we log, for how long, and why no cookie banner is needed. - [Terms of service](https://oden-api.com/terms): Terms for the ODEN API and this website. ## Optional - [Dashboard](https://oden-api.com/app): API key management and usage. Requires sign-in; not useful to crawl.