AI Agent Orchestration

Orchestrate
AI Agents

Define, coordinate, and execute intelligent AI agents with simple YAML configurations. From single agents to complex multi-agent workflows.

Explore Features Request Access

LLM Support

Multi-Provider

Configuration

Declarative YAML

Deployment

Production-Ready

⚡ Multi-Agent

Agent Status

Executing

LLM Provider

Gemini Pro

Scroll to explore

Declarative Simplicity

Define agent behavior, tools, data sources, and LLM configurations using clean, human-readable YAML. No complex code required—just declare what you want your agents to do.

agent:
  name: document-analyzer
  
  system_message: |
    You analyze documents and extract key insights.
  
  tools:
    - FileSystem.read_file
    - WebSearch.google_search
  
  data_type_to_return:
    name: AnalysisResult
    fields:
      - name: summary
        type: str

Structured Outputs

Define exact output schemas for your agents. Get predictable, validated responses every time.

Tool Integration

Equip agents with native tools, MCP servers, external plugins, and custom workspace functions.

Intuitive Interface

Build Agents Visually

No coding required. Create powerful AI agents through our step-by-step wizard with pre-built templates.

Welcome to Weave

Build AI agents that automate tasks, analyze documents, and process information. No coding required.

Quick Start Templates

Choose a template to get started in seconds:

Text Summarizer

Condense long documents into key points

Data Extractor

Pull structured data from unstructured text

Document Analyzer

Analyze and classify documents automatically

Question Answerer

Answer questions based on context

Step 1

Choose Template

Select a starting point or create from scratch

Data Extractor

Document Analyzer

Step 2

Define Behavior

Name your agent and write instructions

                name: data-extractor

                task: Extract key info...

Step 3

Define Output

Set structured fields the agent returns

extracted_data Text

confidence Number

ct_report_analyzer agents_yamls/

Schema

Structured

Raw YAML

agent:
name: "ct_report_analyzer"
system_message: |
You are an expert radiologist AI assistant specializing
              
in brain CT scan analysis for organ donor identification.
              
Analyze the following CT report and
                extract key findings:
1. Temporal hemorrhage (present/absent)
2. Midline shift measurement in millimeters
3. Transtentorial herniation (present/absent)
data_type_to_return:
name: "CTAnalysisResult"
fields:
- name: "temporalHemorrhage"
type: "bool"

Platform Capabilities

Everything you need to build, orchestrate, and deploy intelligent AI agents

Multi-Agent

Agent Coordination

Orchestrate multiple agents working together. Chain them sequentially, pass data between them, and use conditional logic to create sophisticated workflows.

✓ Sequential agent chaining
✓ Automatic data passing between agents
✓ Conditional branching and loops

agent_coordination:
  name: research-pipeline
  
  order:
    - topic-researcher
    - content-drafter
    - content-reviewer
    - final-polisher

Multi-LLM

Any LLM Provider

Use any LLM provider through a unified interface. Switch between providers without changing your agent logic. Run locally with Ollama or scale with cloud providers.

✓ Ollama for local LLMs (Llama, Gemma, Qwen, DeepSeek)
✓ Cloud providers: Gemini, OpenAI, Anthropic
✓ OpenRouter for 100+ models access

Gemini

Anthropic

◬

OpenAI

🦙

Ollama

Python SDK

Simple Integration

Integrate Weave into your Python applications with our intuitive SDK. Run agents, manage workflows, and handle results with just a few lines of code.

✓ Pythonic, intuitive API design
✓ Built-in error handling and retries
✓ File upload support for multi-modal agents

from sdk_python import WeaveClient

client = WeaveClient()

# Run an agent
result = client.run_agent(
    "document-analyzer",
    variables={"input": "Analyze this..."}
)

print(result["final_response"])

How It Works

Agent Lifecycle

From definition to execution in three seamless steps.

Define

Write YAML agents

Orchestrate

Coordinate workflows

Execute

Run with LLM & tools

Hot Reload

Update YAML definitions and deploy instantly. No compilation needed.

Schema Validation

Catch configuration errors before they reach production.

Version Control

Track every change in Git. Roll back configs instantly.

Extensibility

Powerful Tool System

Equip your agents with the tools they need to get work done.

Native Tools

Built-in Python utilities for file ops, web search, and more.

MCP Protocol

Integrate Model Context Protocol servers for external tools.

Plugins

Install third-party tool packages via pip.

Custom Tools

Define workspace tools with @weave_tool decorator.

Ready to Orchestrate?

Start building intelligent multi-agent systems with Weave today.

Request Weave Access

OrchestrateAI Agents