Discovery and streaming generation
POST discovery returns relevant tables, similar queries, and glossary-style context; generation routes stream status, results, and errors over SSE for T-SQL, Python, R, SAS, and a rate-limited code advisor.
The NL-to-SQL & code API
Natural Language to SQL, Streaming Code, and Governed Data Access
Octofy AI Agent is mainly developed to support Octofy Professional at enterprise scale: hosted query generation, interactive debugging, and SQL optimization against your governed data estate. The same FastAPI service is available to analysts and integrators more broadly—semantic discovery from vector metadata, Server-Sent Event streams for SQL and multi-language code generation, and execution endpoints that retry and auto-fix failures. Multi-source APIs cover data connections, schema trees, and object lookup; users sign in for API keys, keep conversation history, and submit contributions while admins sync schemas, curate few-shots and value indexes, verify connectivity, and run backup and skills utilities.
Octofy Pro can talk to any backend that implements the same HTTP contract. The Octofy AI Agent backend API in the Documentation Center lists base URL, authentication, SSE generation, execution, conversations, and admin endpoints.
Type to SQL Sandbox
-- SQL response will stream hereArchitecture Diagram
Click nodes to inspect the integration layer.
Active node: Octofy AI Agent
Data flow: Octofy Pro request -> AI Agent retrieval/rerank -> connected DBMS execution + Milvus memory -> response stream.
Core Features
Execution, profiling, and summaries
Run SQL or Python with optional profiling, chart hints, and multi-source execution; planning summaries and natural-language summaries help interpret previews without leaving the API.
Multi-source schema and navigation
Register and scan data sources, resolve server and database or file paths to stable IDs, browse per-source or combined schema trees, and fetch detailed metadata for specific objects.
Authentication, conversations, and admin
API-key access for users and admins, conversation CRUD for chat history, contribution submit and approve flows, plus administration for schema sync, few-shots, value indexes, settings, models, and vector-store backup.
Release Notes
v1.4.0 - 2026-02-24
- Multi-source data-source and schema-tree APIs aligned with discovery and execution.
- Conversation persistence and user activity endpoints for hosted deployments.
- Admin coverage for schema sync, few-shots, value indexes, and vector-store backup.