Beyond simple
retrieval.
DataMind AI classifies user intent in milliseconds, routing complex queries to specialized agents. Query databases, automate tasks, and fetch documents through a single stream.
Intent Stream
Real-time classification.
Vector Fetch
Instant context retrieval.
Agent Flow
Smart query orchestration.
DataMind at work
Watch how we classify intents and route agents in real-time. No smoke and mirrors, just pure performance.
How it works
One query. The right action.
Every message is classified, routed, and executed automatically — from question to real-world action in milliseconds.
Understanding your query
Intent Classification
The AI router analyses your natural language query in milliseconds, identifying whether you need a database lookup, email automation, spreadsheet analysis, document retrieval, or a multi-step workflow.
Right agent, every time
Smart Agent Routing
Based on classified intent, your query is dispatched to the specialised agent — SQL Agent, Email Agent, CSV/Excel Agent, RAG Knowledge Agent, or Workflow Orchestrator. No manual selection needed.
Real actions, not just text
Action Execution
The agent takes real action — queries your live database, reads your spreadsheet, searches through uploaded documents, or sends that email. Results are computed from your actual data sources.
Answer + proof of action
Structured Response
You receive a natural language answer, the intent classification, the action taken, source citations, and latency — all in one structured JSON response through a single API call or the Playground.
Capabilities
Five agents. One platform.
Each agent is purpose-built for its domain. You never choose — the system routes automatically.
SQL & Database Agent
Natural Language → SQLAsk questions about your database in plain English. The agent generates and executes the SQL, then returns a clean readable answer.
# Natural language → SQL executed automatically
User: "Which customers spent over $5k last month?"
SELECT customer_id, name, SUM(amount) as total
FROM orders
WHERE order_date >= '2024-11-01'
GROUP BY customer_id
HAVING total > 5000
ORDER BY total DESC;Email Workflow Agent
AutomationSend emails, draft messages, and automate recurring reports — all from chat or the API.
CSV & Excel Agent
Spreadsheet AIUpload Excel or CSV files. Ask analytical questions and get human-readable answers backed by the actual data.
"Which region had the highest return rate in Q3?" → West (8.4%), vs avg 5.1%
Knowledge Retrieval Agent
RAGSemantic search over your uploaded documents — PDFs, text, JSON. Returns grounded answers with source citations.
Multi-step Workflow Agent
OrchestrationChain multiple agents together. One query can retrieve data, run SQL, summarise it, and send the result via email — planned and executed automatically.
Playground
Chat with your entire business context
Test your agent before shipping. Ask anything — the system shows which agent handled it, the reasoning steps it took, and the sources it cited.
Streaming SSE responses — see it think in real time
Pipeline steps shown inline — router, agent, result
Citations and confidence scores for every RAG answer
Scoped per workspace — isolate projects cleanly
Live demo — pipeline steps
What was our revenue last quarter?
For developers
One API for every agent
Send a natural language query and get back a structured response — the intent, the action taken, and the answer. The same endpoint handles RAG, SQL, emails, and workflows. Integrate into your product in minutes.
Streaming SSE for real-time token output
Single REST endpoint — all agents behind one URL
Built-in examples in cURL, JavaScript, Python
JWT auth, workspace-scoped, production-ready
# Ask in plain English — the agent figures out the rest
curl -X POST https://api.datamind.ai/query \\
-H "Authorization: Bearer YOUR_TOKEN" \\
-d '{ "question": "Send the Q3 report to Sarah", "workspaceId": "ws_xyz" }'
# Response — action taken, not just text
{
"intent": "email_workflow",
"action": "email_sent",
"to": "sarah@company.com",
"answer": "Report sent to Sarah with Q3 summary.",
"latency_ms": 61
}Want this power in
your own stack?Self-HostedCustom Integration
Due to high infrastructure costs of Vector DBs and LLMs, this live demo is limited. If you're looking to integrate DataMind AI's intent routing into your systems, let's collaborate.