canonical: https://jentic.com/apis/neuralseek.com/neuralseek # NeuralSeek API NeuralSeek is a generative AI platform for retrieval-augmented answers over enterprise knowledge. The API answers natural-language questions with sourced results, orchestrates prompts and agent flows with maiStro, categorizes and extracts text, translates content, and checks API keys and instance configuration. ## For AI agents Answer natural-language questions over a knowledge base, run maiStro prompt and agent flows, categorize and extract text, and translate content. Authenticates with an API key sent in the apikey header. ## Scope Does not host or train the underlying models. Use the NeuralSeek API for retrieval-augmented answers, maiStro flows, categorization, extraction, and translation. ## Capabilities - Answer natural-language questions over a connected knowledge base - Orchestrate prompts and agent flows with maiStro - Categorize text against defined categories - Extract structured data from text - Translate text between languages - Check an API key and instance configuration ## Use cases ### Retrieval-Augmented Q&A An AI agent connected through Jentic can answer questions grounded in enterprise content. The seek operation returns an answer with its supporting sources, so an assistant can respond with cited material instead of an unverified guess. Example prompt: Answer a user's question over the knowledge base and return the supporting sources ### Prompt and Agent Orchestration Complex generative tasks need structured prompting. NeuralSeek's maiStro runs templated prompt and agent flows, letting an agent invoke a defined workflow and read back its generated output rather than assembling prompts by hand. Example prompt: Run a named maiStro template with the supplied inputs and return the output ### Text Classification and Extraction Structured processing of unstructured text is a common step. The API categorizes text against defined categories and extracts structured fields, so an agent can label incoming content and pull out the values a downstream step needs. Example prompt: Categorize a document and extract its key fields ## Key endpoints | Method | Path | Description | | --- | --- | --- | | POST | `/seek` | Answer a question with sourced results | | POST | `/maistro` | Run a maiStro prompt or agent flow | | POST | `/categorize` | Categorize text against defined categories | | POST | `/extract` | Extract structured data from text | | POST | `/translate` | Translate text between languages | ## Key resources - **Seek** — Retrieval-augmented question answering with sources - **maiStro** — Prompt and agent flow orchestration - **Categorize** — Text classification against defined categories - **Extract** — Structured data extraction from text - **Translate** — Text translation between languages ## Why Jentic - **Setup:** Wiring the NeuralSeek API by hand means provisioning an instance key, building seek and maiStro request bodies, and threading the key through every call yourself. Through Jentic you install once, import NeuralSeek from the API Directory, store the key once, and your agent calls the seek and text operations directly. - **Permission scoping:** You choose which of the operations your agent may call. You might allow question answering and translation while blocking configuration and key operations, so the agent can use the platform without changing its setup. - **Credential handling:** Your NeuralSeek 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 'answer a question over the knowledge base', and Jentic returns the matching NeuralSeek operation with its input schema so the agent calls it with the right question. ## Related APIs - **OpenAI API** — Chat completions, embeddings, and other model endpoints - **Cohere** — Language models for generation, embeddings, and reranking - **Hugging Face** — Model hosting and inference across many open models ## FAQ ### What authentication does the NeuralSeek API use? It authenticates with an API key sent in the apikey header, per the OpenAPI spec. Through Jentic the key is stored once, encrypted, by your own self-hosted instance and injected when the agent calls, so it never appears in the agent's prompt or logs. ### Can I limit what my agent is allowed to do with the NeuralSeek API? Yes. You choose which operations the agent may call. You might allow question answering and translation while blocking configuration and key operations, so the agent can use the platform without changing its setup, with every call logged by your own instance. ### Is there a NeuralSeek MCP server? You don't need an MCP server to give your agent the NeuralSeek API. Jentic connects it directly from the API Directory: import it, store the key once, and your agent calls the seek, maiStro, and text operations. Operations are discovered on demand, so nothing extra loads into the agent's context. ### What are the rate limits for the NeuralSeek API? The OpenAPI spec does not define rate limits. Check the NeuralSeek documentation for any plan limits before running high-volume seek or maiStro operations. ### What can the NeuralSeek API do? It answers natural-language questions over a knowledge base with sources, runs maiStro prompt and agent flows, categorizes and extracts text, translates content, and checks API keys and instance configuration. ### How do I answer a question through Jentic? Search Jentic by intent, for example 'answer a question over the knowledge base', and Jentic returns the matching NeuralSeek operation with its input schema. Your agent supplies the question and reads back the sourced answer. To run it on your own infrastructure, install Jentic One from its GitHub repo.