canonical: https://jentic.com/apis/apipie.ai/apipie # APIPie The APIPie API is a single gateway for the model and data operations an agentic workflow needs. It runs chat completions, generates embeddings and images, converts text to speech, and transcribes audio, and it manages vector collections for semantic search. Beyond models it searches and scrapes the web, parses documents with Apache Tika, anonymizes sensitive text, and queries job and real-estate datasets, so an agent reaches many capabilities through one consistent interface. ## For AI agents Call chat, embedding, image, and audio models through one gateway, manage vector collections for semantic search, and run web search, document parsing, and data lookups. ## Scope Does not train or host your own custom models or manage fine-tuning jobs. Use for calling hosted models, search, document parsing, and vector operations only. ## Capabilities - Run chat completions against hosted language models - Generate embeddings and query vector collections for semantic search - Generate images and convert text to speech - Transcribe audio to text - Search and scrape the web for current information - Parse document content and metadata and anonymize sensitive text - Query job and real-estate datasets by keyword or location ## Use cases ### One gateway for an agent's model calls An AI agent connected through Jentic reaches chat, embedding, image, and audio models through APIPie's single interface. It runs a completion, embeds supporting text, and generates an image from one consistent gateway. Because the agent discovers operations by intent, it uses many model capabilities without a developer integrating each provider separately. Example prompt: Run a chat completion for a prompt, generate embeddings for the supporting text, and return both results ### RAG with built-in vectors and parsing APIPie covers the retrieval side of a RAG pipeline: it parses documents with Apache Tika, generates embeddings, and stores and queries vector collections. An agent can ingest a document, embed its chunks, and query the collection for the passages most relevant to a question, assembling retrieval without stitching a separate vector store and parser together. Example prompt: Parse a document, create a vector collection, upsert embeddings for its chunks, and query the collection for a question ### Grounding answers with live web and data search The search, scrape, and data endpoints let an agent ground its answers in current information: it searches the web, scrapes a page, or queries job and real-estate datasets, then reasons over the structured result. This keeps an assistant's responses tied to live sources rather than only its training data. Example prompt: Search the web for a query, scrape the top result, and return the extracted content for the agent to summarize ## Key endpoints | Method | Path | Description | | --- | --- | --- | | POST | `/v1/chat/completions` | Run a chat completion | | POST | `/v1/embeddings` | Generate embeddings | | POST | `/v1/images/generations` | Generate an image | | POST | `/v1/audio/transcriptions` | Transcribe audio to text | | POST | `/v1/search` | Search the web | | POST | `/vectors/query` | Query a vector collection | | GET | `/v1/models` | Fetch available models | ## Key resources - **Chat & Models** — Run chat completions and list available models - **Embeddings & Vectors** — Generate embeddings and manage and query vector collections - **Audio & Images** — Convert text to speech, transcribe audio, and generate images - **Search** — Search and scrape the web and return answers - **Parser & Anon** — Parse document content with Apache Tika and anonymize sensitive text - **Data** — Query job and real-estate datasets ## AI readiness This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use. - **Score:** 57 / 100 - **Maturity:** Foundational - **Dimensions:** - Foundational Compliance: 42 / 100 - Developer Experience & Jentic Compatibility: 72 / 100 - AI-Readiness & Agent Experience: 44 / 100 - Agent Usability: 94 / 100 - Security: 58 / 100 - AI Discoverability: 63 / 100 - **View full report:** https://jentic.com/apis/apipie.ai/apipie/scorecard - **How the score is calculated:** https://docs.jentic.com/reference/api-readiness-framework/overview/ - **More about the dimensions:** https://docs.jentic.com/reference/api-readiness-framework/specification/#dimensional-model-overview ### Score it yourself Every API in the directory is allowlisted, so you can re-score it with no key required. - **Score your own API:** https://jentic.com/scorecard.md - **Scoring CLI agent skill:** https://github.com/jentic/jentic-api-scorecard/blob/main/skills/jentic-api-scorecard/SKILL.md ```sh npx @jentic/api-scorecard-cli score ``` ## Why Jentic - **Setup:** Wiring the APIPie API by hand means handling its header-based API key or bearer auth and learning the request shape for each model and data operation. Through Jentic you install once, import the API from the Directory, store the key once, and your agent calls it. - **Permission scoping:** APIPie carries its inputs in the request body, so rules bound which operations your agent may call: you can allow chat and embedding operations while leaving vector-collection deletion and state changes out unless you add them. You choose the operations it may use. - **Credential handling:** Your APIPie API key is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context. - **Discovery method:** Agents search Jentic by intent such as 'run a chat completion' or 'query a vector collection', and Jentic returns the matching APIPie operation with its input schema so the agent calls the right endpoint without browsing the reference docs. ## Related APIs - **OpenRouter** — OpenRouter is a unified gateway to many language models through one API - **OpenAI** — OpenAI exposes its own chat, embedding, image, and audio models directly - **Cohere** — Cohere provides specialized embedding and rerank models for retrieval ## FAQ ### What authentication does the APIPie API use? The APIPie API authenticates with an API key sent in the Authorization or x-api-key header, or a bearer token, as declared in its OpenAPI spec. Through Jentic the credential is stored once by your own self-hosted instance and injected at call time, so it never appears in the agent's prompt or logs. ### Can I run chat completions and embeddings with the APIPie API? Yes. You call chat completions against hosted models, generate embeddings, and query vector collections for semantic search through the same gateway. It also generates images, converts text to speech, and transcribes audio. ### Is there an APIPie API MCP server? You do not need an MCP server to give your agent the APIPie API. Jentic connects it directly from the API Directory: import the API, store your API key once, and your agent calls models and vector operations on demand without loading another server's tool definitions into its context. ### Can I limit what my agent is allowed to do with the APIPie API? Yes. Write a rule that allows only the chat and embedding operations, so the agent can reason and retrieve but cannot delete vector collections or change state unless you add those, and every call it makes is logged for review. APIPie carries its inputs in the request body, so rules bound which operations the agent may call. ### What are the rate limits for the APIPie API? The OpenAPI spec does not state numeric rate limits for the APIPie API, which is usage-based per the provider. Check the vendor documentation at https://apipie.ai for current limits before running large batch jobs. ### How do I run a chat completion with the APIPie API through Jentic? Search Jentic for an intent such as 'run a chat completion', and it returns the matching operation with its input schema so your agent passes the model and messages correctly. To run it on your own infrastructure, install Jentic One from its GitHub repo.