TiDB AI Assistant

O que é TiDB AI Assistant?

TiDB AI Assistant is a TiDB-focused question-answering product that presents itself as a knowledge-graph-based RAG system built with TiDB X Vector Storage and PyTiDB. The public homepage invites users to ask technical questions about TiDB and related workflows, and the product appears to be exposed through both a web interface and an API.

The examples on the site center on practical TiDB topics such as storing and querying vectors with PyTiDB, using foreign keys, running vector search, full-text search, or hybrid search, and asking about TiDB Cloud pricing comparisons. The API documentation adds more detail by exposing objects for chat, knowledge bases, embeddings, reranking, evaluation, and knowledge-graph retrieval.

O que TiDB AI Assistant pode fazer?

Knowledge-graph-based RAG

The homepage describes the product as a knowledge-graph-based RAG assistant, indicating that it combines retrieval over structured knowledge with generated answers.

TiDB and PyTiDB foundation

The product is built with TiDB X Vector Storage and PyTiDB, which ties the assistant to TiDB-oriented storage and Python access patterns.

API for retrieval and chat workflows

The Swagger UI exposes endpoints and object models for chat, knowledge bases, documents, chunks, entities, relationships, embeddings, reranking, and evaluation tasks.

Multiple retrieval modes

API schema names include vector search, full-text search, hybrid search, and knowledge-graph retrieval, showing multiple retrieval modes in the backend.

Administrative and model management

The docs include admin-oriented objects for API keys, data sources, models, and settings, suggesting the service is configurable through managed resources.

Casos de uso

“TiDB support and reference questions”

Ask operational or conceptual questions about TiDB, including topics such as foreign keys, cloud pricing, and feature behavior.

“Vector and hybrid search exploration”

Query vectors through PyTiDB and compare vector search with full-text or hybrid retrieval approaches.

“Retrieval app prototyping”

Build chat or knowledge-base workflows using the documented API objects for documents, chunks, embeddings, and retrieval.

“Knowledge-graph RAG workflows”

Work with knowledge-graph retrieval for question answering where relationships and entities matter alongside raw text.

“Assistant quality tuning”

Inspect evaluation, reranking, and model-management objects in the API when tuning retrieval quality or assistant behavior.

Perguntas frequentes

How do you access TiDB AI Assistant?

The site shows TiDB AI Assistant as a web app with login access and an API documented through Swagger UI. The public pages do not provide setup steps beyond those entry points.

What does TiDB AI Assistant do?

The homepage presents it as a knowledge-graph-based RAG assistant built with TiDB X Vector Storage and PyTiDB, and the API docs expose knowledge-base, vector search, knowledge-graph retrieval, chat, embedding, reranking, and evaluation-related objects.

Is pricing published?

The rendered pages do not show pricing details. The pricing URL returns a 404, so there is no public pricing information in the collected sources.

What kinds of questions is it aimed at?

The homepage examples focus on TiDB-related questions such as vectors with PyTiDB, foreign keys, vector search, full-text search, hybrid search, and TiDB Cloud cost questions.