No PhD required. We checked.
For developers who value simplicity.

Cloud-agnostic AI agent framework for Python. Build agents that actually do things, coordinate multi-agent teams, and see what's happening inside.

---

Quick Start (60 seconds)

export ANTHROPIC_API_KEY=sk-ant-...
uvx bedsheet  # Run demo instantly, no install needed

A research assistant in 20 lines:

import asyncio
from bedsheet import Agent, ActionGroup
from bedsheet.llm import AnthropicClient
from bedsheet.events import CompletionEvent

# Give your agent a superpower
tools = ActionGroup(name="Research")

@tools.action(name="search", description="Search for information")
async def search(query: str) -> str:
    # Your real implementation here (API calls, database, etc.)
    return f"Found 3 results for '{query}': ..."

# Create the agent
agent = Agent(
    name="Researcher",
    instruction="You help users find information. Use the search tool.",
    model_client=AnthropicClient(),
)
agent.add_action_group(tools)

# That's it. Use it.
async def main():
    async for event in agent.invoke("session-1", "What's new in Python 3.12?"):
        if isinstance(event, CompletionEvent):
            print(event.response)

asyncio.run(main())

Want the fancy demo?

uvx bedsheet  # Multi-agent investment advisor with parallel execution

📺 See demo output (click to expand)

============================================================
  BEDSHEET AGENTS - Investment Advisor Demo
============================================================

User: Analyze NVIDIA stock for me

[3.9s] PARALLEL DELEGATION - dispatching 2 agents:
        -> MarketAnalyst: Analyze NVDA stock data and technicals
        -> NewsResearcher: Find and analyze news about NVIDIA

[18.2s] || [MarketAnalyst] Starting...
        [MarketAnalyst] -> get_stock_data({'symbol': 'NVDA'})
        [MarketAnalyst] -> get_technical_analysis({'symbol': 'NVDA'})
        [MarketAnalyst] <- {'symbol': 'NVDA', 'price': 875.5, ...}

[18.2s] || [NewsResearcher] Starting...
        [NewsResearcher] -> search_news({'query': 'NVIDIA'})
        [NewsResearcher] -> analyze_sentiment({'articles': [...]})

[18.2s] OK [MarketAnalyst] Complete
[18.2s] OK [NewsResearcher] Complete

FINAL RESPONSE (32.3s)
------------------------------------------------------------
# NVIDIA (NVDA) Comprehensive Stock Analysis

## Executive Summary
NVIDIA shows **strong bullish signals** across both technical
indicators and fundamental news sentiment...

---

Why "Bedsheet"?

A playful jab at AWS Bedrock Agents. We "cover" the same concepts (agents, action groups, orchestration) but you define everything in code, not through a web console with 15 screens and a 3-minute deployment cycle.

Like a bedsheet fits any bed regardless of brand, Bedsheet fits any cloud—or no cloud at all.

Also, agent frameworks shouldn't take themselves too seriously. The robots aren't sentient yet.

---

The Problem

After years of building with existing frameworks:

Framework	Experience
LangChain	400 pages of docs. Still confused. "Hello world" = 47 lines.
AWS Bedrock	Click. Wait. Click. Wait. Change one word. Repeat for eternity.
AutoGPT	Agent "researched" by opening 200 browser tabs. RIP laptop.
CrewAI	2 hours configuring "crew dynamics". Agents still fighting.

Bedsheet's philosophy:

# This is the entire mental model
async for event in agent.invoke(session_id, user_input):
    print(event)  # See everything. Debug anything. Trust nothing.

---

Features

Single Agent + Tools

tools = ActionGroup(name="Math")

@tools.action(name="calculate", description="Do math")
async def calculate(expression: str) -> float:
    return eval(expression)  # Don't actually do this in production

agent = Agent(
    name="Calculator",
    instruction="Help with math. Use the calculate tool.",
    model_client=AnthropicClient(),
)
agent.add_action_group(tools)

Multi-Agent Teams

The good stuff. A Supervisor coordinates specialized agents:

from bedsheet import Supervisor

researcher = Agent(name="Researcher", instruction="Research topics.", ...)
writer = Agent(name="Writer", instruction="Write clearly.", ...)

supervisor = Supervisor(
    name="ContentTeam",
    instruction="""Coordinate content creation:
    1. Have Researcher gather info
    2. Have Writer create the piece
    Synthesize the final result.""",
    model_client=AnthropicClient(),
    collaborators=[researcher, writer],
)

Parallel Execution

Why wait for agents one-by-one?

# In supervisor instruction:
# "Delegate to BOTH agents simultaneously..."

delegate(delegations=[
    {"agent_name": "MarketAnalyst", "task": "Get stock data"},
    {"agent_name": "NewsResearcher", "task": "Find news"}
])

# Both run at the same time
# Sequential: 4 seconds → Parallel: 2 seconds

Event Streaming

See everything happening inside:

async for event in agent.invoke(session_id, user_input):
    match event:
        case ToolCallEvent(tool_name=name):
            print(f"Calling: {name}")
        case DelegationEvent(delegations=d):
            print(f"Delegating to: {[x['agent_name'] for x in d]}")
        case CompletionEvent(response=r):
            print(f"Done: {r}")
        case ErrorEvent(error=e):
            print(f"Oops: {e}")  # At least you know what broke

Two Modes

Mode	What It Does	Use When
supervisor	Coordinates agents, synthesizes results	Complex tasks
router	Picks one agent, hands off completely	Simple routing

Structured Outputs (v0.3+)

Guarantee your agent returns valid JSON matching your schema. Uses Anthropic's native constrained decoding—the model literally cannot produce invalid output.

from bedsheet.llm import AnthropicClient, OutputSchema

# Option 1: Raw JSON schema (no dependencies)
schema = OutputSchema.from_dict({
    "type": "object",
    "properties": {
        "symbol": {"type": "string"},
        "recommendation": {"type": "string", "enum": ["buy", "sell", "hold"]},
        "confidence": {"type": "number", "minimum": 0, "maximum": 1},
    },
    "required": ["symbol", "recommendation", "confidence"]
})

# Option 2: Pydantic model (if you prefer)
from pydantic import BaseModel

class StockAnalysis(BaseModel):
    symbol: str
    recommendation: str
    confidence: float

schema = OutputSchema.from_pydantic(StockAnalysis)

# Use with any LLM call
client = AnthropicClient()
response = await client.chat(
    messages=[{"role": "user", "content": "Analyze NVDA"}],
    system="You are a stock analyst.",
    output_schema=schema,  # 100% guaranteed valid JSON
)

# Access the validated data
print(response.parsed_output)  # {"symbol": "NVDA", "recommendation": "buy", "confidence": 0.85}

Key points:

• ✅ Works WITH tools (unlike Google ADK which disables tools with schemas)
• ✅ Pydantic is optional—raw JSON schemas work fine
• ✅ Uses Anthropic's beta structured-outputs-2025-11-13 under the hood
• ✅ Zero chance of malformed JSON—constrained at token generation

---

Real Example: Todo Assistant

Something actually useful:

import asyncio
from bedsheet import Agent, ActionGroup
from bedsheet.llm import AnthropicClient
from bedsheet.events import CompletionEvent, ToolCallEvent

todos = []  # Use a real database

tools = ActionGroup(name="Todos")

@tools.action(name="add_todo", description="Add a todo item")
async def add_todo(task: str, priority: str = "medium") -> dict:
    todo = {"id": len(todos) + 1, "task": task, "priority": priority, "done": False}
    todos.append(todo)
    return todo

@tools.action(name="list_todos", description="List all todos")
async def list_todos() -> list:
    return todos

@tools.action(name="complete_todo", description="Mark todo as done")
async def complete_todo(todo_id: int) -> dict:
    for t in todos:
        if t["id"] == todo_id:
            t["done"] = True
            return t
    return {"error": "Not found"}

assistant = Agent(
    name="TodoBot",
    instruction="Manage the user's todo list. Be helpful and concise.",
    model_client=AnthropicClient(),
)
assistant.add_action_group(tools)

async def main():
    queries = [
        "Add a task: Buy milk",
        "Add: Call mom, high priority",
        "What's on my list?",
        "Done with the milk!",
    ]
    for q in queries:
        print(f"\nYou: {q}")
        async for event in assistant.invoke("user-1", q):
            if isinstance(event, CompletionEvent):
                print(f"Bot: {event.response}")

asyncio.run(main())

---

Installation

# Recommended: Use uv for fast, reliable installs
uv pip install bedsheet           # Basic
uv pip install bedsheet[redis]    # + Redis memory backend
uv pip install bedsheet[dev]      # + Development tools

# Or run directly without installing
uvx bedsheet --help

Requirements: Python 3.11+ and an Anthropic API key

---

Architecture

bedsheet/
├── agent.py          # Single agent (189 lines)
├── supervisor.py     # Multi-agent coordination (362 lines)
├── action_group.py   # Tool definitions (115 lines)
├── events.py         # Event types (105 lines)
├── llm/
│   ├── base.py       # LLM protocol
│   └── anthropic.py  # Claude implementation
└── memory/
    ├── in_memory.py  # Development
    └── redis.py      # Production

Total: ~1,000 lines. Coffee break reading.

---

Comparison

	Bedsheet	LangChain	AWS Bedrock	CrewAI
Lines of code	~1,000	~100,000+	N/A	~10,000
Time to understand	1 afternoon	1 week	2 days	3 days
Debugging	print() works	Good luck	CloudWatch	Logs
Streaming events	Built-in	Add-on	Limited	Limited
Parallel execution	Default	Manual	Manual	Manual
Cloud lock-in	None	None	AWS	None

---

Documentation

• User Guide — Beginner to advanced, 12 lessons
• Technical Guide — Python patterns explained
• Deployment Guide — Local, GCP, and AWS deployment
• Multi-Agent Guide — Supervisor deep dive

---

Roadmap

• v0.1 — Single agents, tools, streaming
• v0.2 — Multi-agent, parallel delegation
• v0.3 — Structured outputs
• v0.4 — Deploy anywhere (Local/GCP/AWS), Debug UI ← you are here
• v0.5 — Knowledge bases, RAG
• v0.6 — Guardrails, safety
• v0.7 — GCP Agent Engine, A2A protocol

---

Contributing

git clone https://github.com/sivang/bedsheet.git
cd bedsheet
uv pip install -e ".[dev]"
pytest -v  # 180 tests, all green

See CONTRIBUTING.md for guidelines.

---

FAQ

Production ready? Yes. 180 tests, type hints, async-first, Redis support. We use it.

Only Claude? For now. LLMClient is a protocol—implement it for OpenAI/Gemini/local. PRs welcome.

Why not LangChain? Life is short.

Is the name a joke? Yes. The code isn't.

---

License

Elastic License 2.0 - see LICENSE for details.

---

Copyright © 2025-2026 Sivan Grünberg, Vitakka Consulting

_{Star if it helped. Issue if it didn't. Either way, we're listening.}