crewai

CrewAI + Moss Cookbook Example

Use Moss semantic search as a retrieval tool for CrewAI agents. Agents get sub-10ms search over your knowledge base.

Installation

pip install crewai moss python-dotenv

Setup

Set your credentials as environment variables or in a .env file (see .env.example):

MOSS_PROJECT_ID=your-project-id
MOSS_PROJECT_KEY=your-project-key
GEMINI_API_KEY=your-gemini-key

Quick Start

from crewai import Agent, Task, Crew
from moss import MossClient
from moss_crewai import MossSearchTool

client = MossClient("your-project-id", "your-project-key")

search = MossSearchTool(client=client, index_name="knowledge-base", top_k=5)

researcher = Agent(
    role="Research Assistant",
    goal="Find accurate answers using the knowledge base",
    tools=[search],
)

task = Task(
    description="What are the best budget hotels in Tokyo?",
    expected_output="A concise answer based on the knowledge base.",
    agent=researcher,
)

crew = Crew(agents=[researcher], tasks=[task])
result = crew.kickoff()

Demo: Multi-Agent Travel Planner

The included example_usage.py runs an interactive CLI travel planner with 4 agents:

User Question (e.g. "Plan a 2-day trip to Tokyo on a budget")
     |
     v
+-------------------+  +------------------+  +-------------------+
| Destinations      |  | Hotels & Stays   |  | Activities &      |
| Specialist        |  | Specialist       |  | Tours Specialist  |
| (travel-          |  | (travel-stays)   |  | (travel-          |
|  destinations)    |  |                  |  |  activities)      |
+--------+----------+  +--------+---------+  +---------+---------+
         |                      |                      |
         +----------------------+----------------------+
                                |
                                v
                      +------------------+
                      | Travel Planner   |
                      | (synthesize)     |
                      +------------------+
                                |
                                v
                          Final Answer

How it works

1. Setup: Travel data is indexed into 3 Moss indexes:

   • travel-destinations (10 docs) — city guides, budget tips, transport, best times to visit
   • travel-stays (11 docs) — hotels, hostels, prices, amenities
   • travel-activities (24 docs) — tours, sightseeing, dining, experiences with costs

2. 3 Specialist Agents each search their Moss index using MossSearchTool

3. Travel Planner Agent synthesizes all findings into an actionable travel plan

Run the demo

cd examples/cookbook/crewai
python example_usage.py

=== Moss + CrewAI Travel Planner ===
Plan your next trip! Type 'quit' to exit.

You: Plan a 2-day trip to Tokyo on a budget
Agent: ...

Available Tools

All tools are defined in moss_crewai.py. Copy this file into your project and import from it.

Search

Tool	Agent Input	MossClient Method	Description
MossSearchTool	query: str	query()	Semantic search with ranked results and relevance scores

Configuration (set at construction, not controlled by the agent):

Parameter	Default	Description
client	required	Shared MossClient instance
index_name	required	Index to search
top_k	5	Number of results
alpha	0.8	Hybrid search balance (0=keyword, 1=semantic)

Document Management

Tool	Agent Input	MossClient Method	Description
MossAddDocsTool	texts: list[str], ids?: list[str], upsert?: bool	add_docs()	Add documents to an index. Set upsert=True to update existing docs.
MossDeleteDocsTool	doc_ids: list[str]	delete_docs()	Delete specific documents by their IDs
MossGetDocsTool	doc_ids?: list[str]	get_docs()	Retrieve documents. Fetches all if no IDs provided.

Index Management

Tool	Agent Input	MossClient Method	Description
MossCreateIndexTool	index_name: str, texts: list[str], ids?: list[str]	create_index()	Create a new index populated with documents
MossDeleteIndexTool	index_name: str	delete_index()	Delete an index and all its data (irreversible)
MossGetIndexTool	index_name: str	get_index()	Get info about a specific index (doc count, status)
MossListIndexesTool	(none)	list_indexes()	List all indexes with doc counts and status

moss_tools() Factory

Create all 8 tools with shared configuration:

from moss import MossClient
from moss_crewai import moss_tools

client = MossClient("your-project-id", "your-project-key")
tools = moss_tools(client=client, index_name="knowledge-base")

Files

File	Description
moss_crewai.py	8 tool classes + moss_tools() factory
example_usage.py	Multi-agent travel planner CLI demo with 3 specialists and writer
data/	Travel data: destinations_moss.json, stays_moss.json, activities_moss.json
test_live.py	Live platform tests against real Moss API
.env.example	Template for required environment variables