Self-learning Retrieval for Agents

Smart agents need needles, not haystacks.

More context is more hay, whether in the context window or the data warehouse. Antfly's self-learning retrieval finds answers in the data you already have. Query your own data in minutes.

Run your first query See What You Can Build

Quickstart

terminal

# Install and start in swarm mode (single-node, free)
curl -fsSL https://releases.antfly.io/antfly/latest/install.sh | sh
antfly swarm

# Create a table with a semantic index (embedded locally, no API keys)
curl -X POST http://localhost:8080/api/v1/tables/docs \
  -H "Content-Type: application/json" \
  -d '{
    "indexes": {
      "content_semantic": {
        "type": "embeddings",
        "template": "{{content}}",
        "embedder": {"provider": "termite", "model": "BAAI/bge-small-en-v1.5"}
      }
    }
  }'

# Insert a document (chunked and embedded automatically)
curl -X POST http://localhost:8080/api/v1/tables/docs/batch \
  -H "Content-Type: application/json" \
  -d '{"inserts":{"intro":{"content":"An agent believes what it retrieves."}}}'

# Query by meaning, with BM25 keyword matching fused in
curl -X POST http://localhost:8080/api/v1/query \
  -H "Content-Type: application/json" \
  -d '{
    "table": "docs",
    "semantic_search": "what does an agent trust?",
    "full_text_search": {"query": "content:retrieves"},
    "indexes": ["content_semantic"],
    "fields": ["content"],
    "limit": 5
  }'

Keep your context layer

You bring the context. Antfly harvests it.

Data warehouses, storage buckets, PDFs, tickets: your context layer is already rich with answers. Antfly runs self-learning retrieval over it, right where it lives, and hands your agents what they need. Nothing to migrate, nothing to replace.

Ready to scale?

Antfly Cloud

Multi-node, managed, and monitored. The same Antfly you run locally, deployed and scaled for you.

View Pricing