13F holdings, fund holdings. The latest 13F-HR holdings of an institutional investment manager (hedge fund, mutual fund family, bank) from SEC EDGAR: each position's issuer, class, CUSIP, value in USD, shares and portfolio weight, largest first, plus the report period, total value and position count. Give the filer by name (e.g. Berkshire Hathaway) or CIK. Public SEC data, filed quarterly up to 45 days after the quarter. Use to see what a fund owns.

Web research. One call to research a question on the web: a Brave web search, then the text of the top 1-5 result pages, read in parallel and clipped to your length, with titles and URLs. A page that cannot be read is listed with its reason. Results are for the caller's immediate use only: no storing, resale or model training (Brave API terms). Use when an agent needs current, sourced information in one step instead of a search plus several page reads.

URL safety check, scam signals. Check a link before an agent opens or pays it: where it really leads (each redirect hop, the final URL), the destination domain's age, registrar and trust flags, HTTPS and security headers, and named risk signals (days-old domain, bare IP host, punycode look-alike, user@host trick, chains across many hosts). Heuristic signals from public facts, not a phishing blocklist. Use to vet links in emails, chats and scraped pages, or a payee's website.

News sentiment. The news on a topic and its tone in one call: the latest headlines from the Brave News index (title, URL, publisher, age), each with a positive or negative score from a DistilBERT model, and the balance across them. Up to 10 articles from the past day, week or month. Not a market signal. Results are for the caller's immediate use only (Brave API terms). Use for a market mood check, brand monitoring or a news brief.

Federal Register search. Search the US Federal Register: final rules, proposed rules, notices and presidential documents, newest first, with title, type, agencies, abstract, publication date, comment deadline, effective date and links. Filter by type and agency. From the Federal Register API (US government work). Use to track new regulations or open comment periods on a topic.

Rerank documents by relevance. Order up to 50 documents by how well each answers a query, with a relevance score from 0 to 1 (BAAI bge-reranker-base, a cross-encoder). Send documents as a JSON array or one per line; each is read up to 2,000 characters, 40,000 in all. Returns the original index, score and text, best first; top_n keeps only the best. Also available in the Cohere-style shape at /v1/rerank. Use to pick what a model reads after search or vector retrieval.

Agent memory, delete a key. Delete the value saved under a key in your namespace now, before it expires. Answers deleted: false when there was nothing to delete. Values are kept until they expire (at most 30 days) or are deleted; reads anywhere may lag a write made elsewhere by up to a minute. Use to forget a fact or clear task state.

Page contents API (urls or ids, text.maxCharacters), pay per request: clean text, title, author and date of up to 5 pages in one call. No account or API key; failed calls are not charged.

OpenAI-compatible embeddings, pay per request: one input string in, a 1,024-dimension BGE-M3 vector out (multilingual, for search and RAG). No account or API key; failed calls are not charged.

Agent memory, save. Save a value (up to 8,000 characters of text or JSON) under a key in your own namespace, so an agent can remember facts, preferences or task state between runs. Overwrites an existing key; expires after 1 to 30 days (default 30). Values are kept until they expire (at most 30 days) or are deleted; reads anywhere may lag a write made elsewhere by up to a minute. Use to persist what an agent learned for a later call to memory_get.

OpenAI-compatible image generation, pay per image: a prompt in, one 1024x1024 FLUX.1 schnell JPEG out as b64_json. No account or API key; failed calls are not charged.

Mistral Small 3.1 chat API. Chat with Mistral AI's Mistral Small 3.1 24B model (Apache 2.0) through one paid HTTP call: send a prompt and an optional system prompt, get the answer, up to 1,024 output tokens. Also reachable as model "mistral-small" on the OpenAI-compatible /v1/chat/completions. No account or provider key. Use when an agent wants this model by name.

Qwen3 30B chat API. Chat with Alibaba's Qwen3 30B-A3B mixture-of-experts model (Apache 2.0) through one paid HTTP call: send a prompt and an optional system prompt, get the answer, up to 1,024 output tokens; its reasoning is removed from the answer. Also reachable as model "qwen3-30b" on the OpenAI-compatible /v1/chat/completions. No account or provider key. Use when an agent wants this model by name.

Agent memory, list keys. List up to 100 keys saved in your namespace, optionally only those starting with a prefix, with their size, save time and expiry; pass the returned cursor for the next page. Values are not included (read them with /memory/get). Values are kept until they expire (at most 30 days) or are deleted; reads anywhere may lag a write made elsewhere by up to a minute. Use to see what an agent has saved.

Content moderation, safety check. Classify a text as safe or unsafe with Meta's Llama Guard 3 and name the hazard categories it falls into (violent and non-violent crimes, sexual content, child exploitation, hate, self-harm, weapons, privacy, defamation, elections and others; 14 in all). Up to 8,000 characters. Also available in the OpenAI moderation shape at /v1/moderations. Use to screen user messages, model outputs or content before an agent posts or acts on it.

Agent memory, recall. Read the value saved under a key in your namespace with /memory/set, with when it was saved and when it expires. A key that was never saved or has expired is a 404 and is not charged. Values are kept until they expire (at most 30 days) or are deleted; reads anywhere may lag a write made elsewhere by up to a minute. Use to recall what an agent saved in an earlier run.

Qwen2.5 Coder 32B, code LLM API. Chat with Alibaba's Qwen2.5 Coder 32B code model (Apache 2.0) through one paid HTTP call: send a prompt and an optional system prompt, get the answer, up to 1,024 output tokens. Also reachable as model "qwen-coder" on the OpenAI-compatible /v1/chat/completions. No account or provider key. Use when an agent wants this model by name.

GPT-OSS 20B chat API. Chat with OpenAI's open-weight gpt-oss-20b reasoning model (Apache 2.0) through one paid HTTP call: send a prompt and an optional system prompt, get the answer, up to 1,024 output tokens. Also reachable as model "gpt-oss-20b" on the OpenAI-compatible /v1/chat/completions. No account or provider key. Use when an agent wants this model by name.

Whale alert, large transfers. Large stablecoin transfers in the last minutes, read live from the network: every USDC and USDT transfer of at least min USD (default 100,000) on Ethereum, Base, Arbitrum, Optimism or Polygon, largest first, with sender, receiver, transaction and block, mints and burns marked. Covers up to 15 minutes on Ethereum and 3 on Base (busiest). Addresses are not labelled. Use to watch whale flows, treasury moves or stablecoin mints.

x402engine — OpenAI's cheapest GPT-5 tier — ultra-affordable for high-volume agent tasks