Compact chat on Granite 4.0 H Micro (@cf/ibm-granite/granite-4.0-h-micro). Lowest floor cost for simple agent loops and classification-style prompts.

Low-latency chat on Llama 3.2 1B Instruct (@cf/meta/llama-3.2-1b-instruct). Best for quick replies and lightweight agents; billed per neuron usage.

Generate text embeddings with BGE Base EN V1.5 (@cf/baai/bge-base-en-v1.5). POST text or texts[] for semantic search, RAG retrieval, and clustering; billed by input character volume.

Higher-quality chat on Llama 3.1 8B Instruct Fast (@cf/meta/llama-3.1-8b-instruct-fast). POST messages[]; suited for reasoning-heavy turns with dynamic token pricing.

Generate text embeddings with BGE Small EN V1.5 (@cf/baai/bge-small-en-v1.5). POST text or texts[] for semantic search, RAG retrieval, and clustering; billed by input character volume.

Generate text embeddings with BGE Large EN V1.5 (@cf/baai/bge-large-en-v1.5). POST text or texts[] for semantic search, RAG retrieval, and clustering; billed by input character volume.

Translate text with M2m100 1.2B (@cf/meta/m2m100-1.2b). POST source text and target language; returns translated_text with dynamic token pricing.

Translate text with Indictrans2 EN Indic 1B (@cf/ai4bharat/indictrans2-en-indic-1B). POST source text and target language; returns translated_text with dynamic token pricing.

Classify sentiment (positive/negative) with DistilBERT on Workers AI. POST text; returns label and confidence score. Fixed low per-request price.

Transcribe speech to text with Whisper (@cf/openai/whisper). POST audio (base64) or audio URL; returns transcript text billed by audio duration.

Summarize long text with Llama 3.2 3B Instruct (@cf/meta/llama-3.2-3b-instruct). POST input text; returns a concise summary. Price scales with input and output length.

Batch upsert up to 64 documents into a namespace. One paid request for agent memory ingest.

Distinct vessels with at least one AIS ping inside a bounding box (`min_lat`/`min_lon`/`max_lat`/`max_lon` in WGS84) over the last `hours` hours. Page size locked at 10 vessels per paid call. Each row carries identity + last ping (`last_lat`, `last_lon`, `last_sog`, `last_cog`, `last_ping_utc`, `destination`) + the sanctioned/watchlisted booleans. Optional facet filters mirror /vessels/search: `flag`, `type`, `subtype`, `commercial_market_name`, `sanctioned`, `watchlisted`. Pagination is…

POST /v1/rerank — @cf/baai/bge-reranker-base (Workers AI, x402)

POST /v1/embed/small — @cf/baai/bge-small-en-v1.5 (Workers AI, x402)

POST /v1/llm/glm-5.2 — @cf/zai-org/glm-5.2 (Workers AI, x402)

POST /v1/chat/micro — @cf/ibm-granite/granite-4.0-h-micro (Workers AI, x402)

POST /v1/guard/llama-guard-3-8b — @cf/meta/llama-guard-3-8b (Workers AI, x402)

HYRE Agent — LP strategy advisor: optimal Meteora DLMM range + capital allocation with AI break-even projection.

HYRE Agent — natural language DeFi query. Ask anything about Solana + EVM DeFi and get LLM-orchestrated answer with signal + confidence + sources.