agno-moss 0.0.1

Moss semantic search integration for Agno agents

agno-moss

The Moss in-memory semantic search runtime for Agno agents.

Moss manages embeddings internally and serves queries from an in-memory runtime — sub-10ms lookups, no external embedder, no vector database to run. Point Knowledge at MossRuntime and Agno agents get instant RAG with zero infrastructure.

Installation

pip install agno-moss
# or
uv add agno-moss

Prerequisites

• Moss project ID and project key — get them from the Moss Portal
• Python 3.10+
• An Agno-compatible model provider (OpenAI, Anthropic, etc.)

Quickstart

import os
from agno.agent import Agent
from agno.knowledge.knowledge import Knowledge
from agno.models.anthropic import Claude
from agno_moss import MossRuntime

knowledge = Knowledge(
    vector_db=MossRuntime(
        index_name="my-index",
        # Falls back to MOSS_PROJECT_ID / MOSS_PROJECT_KEY env vars
    ),
)

agent = Agent(
    model=Claude(id="claude-sonnet-4-20250514"),
    knowledge=knowledge,
    search_knowledge=True,
    markdown=True,
)

knowledge.load(recreate=False)
agent.print_response("What do you know about our return policy?", stream=True)

Configuration

MossRuntime

Parameter	Default	Description
index_name	(required)	Name of the Moss index
project_id	MOSS_PROJECT_ID env var	Moss project ID
project_key	MOSS_PROJECT_KEY env var	Moss project key
embedding_model	"moss-minilm"	"moss-minilm" (fast) or "moss-mediumlm" (higher accuracy)
alpha	0.8	Hybrid search blend: 1.0 = semantic only, 0.0 = keyword only
auto_refresh	False	Auto-refresh the in-memory index when new docs are added
polling_interval_in_seconds	600	Refresh interval when auto_refresh=True

How it works

MossRuntime implements Agno's VectorDb base class:

• create() — loads an existing index into Moss's in-memory runtime. Call once at startup for fast first queries.
• upsert() — creates the index on first call, then adds or updates documents. Loads the index automatically after each batch.
• search() — hybrid semantic + keyword search via the loaded in-memory runtime. Falls back to the cloud API if the index is not loaded.

Moss filters metadata only when the index is loaded locally. content_hash_exists() returns False when unloaded (safe: forces re-upsert rather than silently skipping).

Choosing a model provider

# OpenAI
from agno.models.openai import OpenAIChat
agent = Agent(model=OpenAIChat(id="gpt-4o"), knowledge=knowledge, search_knowledge=True)

# Anthropic
from agno.models.anthropic import Claude
agent = Agent(model=Claude(id="claude-sonnet-4-20250514"), knowledge=knowledge, search_knowledge=True)

See the Agno model providers docs for the full list.

License

BSD 2-Clause — see LICENSE.

Support

• Moss Docs
• Agno Docs