Practical overview
When it fits
Use Datamind when an agent needs to write inspectable state while running and draw on that state in later questions, especially when the workflow spans documents, structured databases, graphs, reusable skills or scoped memory.
Who it is for
- Developers adding writable, persistent information to agent workflows.
- Teams building retrieval workflows that need evidence alongside answers.
- Python developers combining document search, SQL, graph knowledge, reusable skills and scoped memory.
Documented uses
- Store a fact, document, CSV row, graph relationship, skill or memory during an interaction.
- Retrieve information added during the conversation in response to a later question.
- Query structured data with SQL and return table evidence.
- Search documents semantically, traverse graph relationships and recall scoped profile or session memories.
Requirements & limits
The project is publicly available under the Apache-2.0 license. The README documents installation through pip and local use; the quick-start chat configuration requires an LLM API base, key, protocol and model.
- Default storage and retrieval providers include Chroma with BM25, SQLAlchemy, NetworkX, profile-scoped SKILL.md files and SQLite memory.
- Memory supports global, profile and session scopes.
- The built-in native loop is identified as the stable default; the SDK loop is optional and adds Claude Agent SDK features.
- The README describes SQLite profile storage as a local single-user baseline and recommends a shared database and storage backend for multi-process deployment.
- The Codex integration is a thin MCP adapter that shares Datamind’s runtime profile, configuration and storage model.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Set up Datamind for a local chat session
Install the library, configure an LLM connection and start the documented chat interface.
Before you start
- Python environment capable of installing the datamind package
- LLM API base, API key, protocol and model
Steps
Install the package with pip install datamind.
Set DATAMIND__LLM__API_BASE, DATAMIND__LLM__API_KEY, DATAMIND__LLM__PROTOCOL and DATAMIND__LLM__MODEL.
Run datamind chat.
Expected outcome
The documented command starts a Datamind chat session using the configured LLM connection.
Guide limitations
- The documented quick start requires an external LLM connection configuration.
- This guide covers the chat start path, not a shared or public deployment.
Start the local Datamind UI
Launch the documented local FastAPI interface.
Before you start
- Datamind installed in a Python environment
Steps
Run python -m uvicorn datamind.server:app --port 8000.
Open http://127.0.0.1:8000.
Expected outcome
The README describes the local interface as available at the stated address after starting the server.
Guide limitations
- The local SQLite profile is described as a single-user baseline.
- Public or team deployments require additional authentication, authorization, TLS, rate limits and an appropriate shared backend.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 14 October 2026
- Original website title
- OpenDCAI/DataMind · Official homepage title · retained in its source language