nucliadb-agentic-api 1.0.0.post156

NucliaDB Agentic API for Hyperforge

NucliaDB Agentic API

The NucliaDB Agentic API package exposes NucliaDB-oriented agentic capabilities for Hyperforge, including ASK/search flows and MCP integrations.

Install

From the workspace root:

uv sync

Run

Start the service:

uv run nucliadb-agentic-api

Useful endpoints:

• /health/ready
• /health/alive
• /metrics

Configuration

Runtime configuration is provided through environment variables consumed by Pydantic settings. Common settings include:

• HTTP_HOST and HTTP_PORT
• MEMORY_READER_NUCLIADB, MEMORY_WRITER_NUCLIADB, and MEMORY_SEARCH_NUCLIADB
• MEMORY_APIKEY_NUCLIADB
• VALKEY_URL
• IDP_REGIONAL_GRPC
• LOAD_MODULES

Development

Run the package tests from the workspace root:

make test

Objective

Enhance the RAG experience by offering RAO features directly in a KB.

How?

The same way we store search configs in the KB, we could store agentic configs, and when calling the /ask endpoint we can refer to a given agentic config.

Scope

The corresponding RAO workflow will be:

• a Rephrase (optionally)
• a SmartAgent
• a Summarize

Possible sources for the SmartAgent:

• The current KB (possibly several times with different filters)
• Sync service (using connections defined in the current KB)
• Google
• Perplexity
• MCP

Out of scope

• Other KBs
• SQL
• Snowflake (?)
• Sitefinity (?)

Config spec

{
  "rephrase": { // optional
    "ask_to": <filter_expression>, //optional
    "prompt": <string>, // optional
    "model": <model>, // optional
  },
  "smart_agent": {
    "mode": <reactive | plan_execute>,
    "extra_prompt": <string>, // optional
    "models": { // optional
      "context_validation": <model>, // optional
      "planner": <model>, // optional
      "executor": <model>, // optional
    },
    "sources": [<source list>]
  },
  "summarize": { // optional
    "user_prompt": <string>, // optional
    "system_prompt": <string>, // optional
    "conversational": <boolean>,
    "model": <model>, // optional
    // and citations must be forced to chunk-level
  }
}

And sources can be:

{
  "type": "nucliadb",
  "description": <string>,
  "filter_expression": <filter_expression>, //optional
}

{
  "type": "sync",
  "description": <string>,
  "connection": <connection_id>
}

{
  "type": "google",
  "description": <string>
}

{
  "type": "perplexity",
  "description": <string>
}

{
  "type": "mcp",
  "description": <string>,
  <...the MCP driver params>
}