3 lines of code.
Infinite actions for any LLM.
Toolhouse is the first cloud infrastructure for tool use.
Instantly give advanced actions and knowledge to your AI apps and agents.
Our curated library of LLM tools integrates with a single click.
Works across any framework and LLM.
# Anthropic th = Toolhouse() client = Anthropic() messages = [{ "role": "user", "content": "Generate code to calculate the Fibonacci sequence to 100." "Execute it and give me the result." }] response = client.messages.create( max_tokens=1024, model="claude-3-5-sonnet-20240620", messages=messages, # Provides a sandboxed code interpreter tools=th.get_tools(), ) messages += th.run_tools(response)# OpenAI Completions th = Toolhouse() client = OpenAI() messages = [{ "role": "user", "content": "Look at my Linear backlog and send a standup update on Slack." }] response = client.chat.completions.create( max_tokens=1024, model="gpt-4o", messages=messages, # Provides Linear and Slack API integrations tools=th.get_tools(), ) messages += th.run_tools(response)# Llama on Groq th = Toolhouse() client = Groq() messages = [{ "role": "user", "content": "Get the temperature record for San Francisco," "compute the median, and give me the result." }] response = client.chat.completions.create( model="llama3-70b-8192", messages=messages, # Provides weather API and a sandboxed code interpreter tools=th.get_tools(), ) messages += th.run_tools(response)# Command-r th = Toolhouse() client = cohere.Client() message = "I'm hungry! Can you give me a recipe for one of my favorite food?" response = client.chat( model="command-r", message=message, # Provides a memory tool tools=th.get_tools(), ) tool_results = th.run_tools(response)# Anthropic th = Toolhouse() client = Anthropic() messages = [{ "role": "user", "content": "Generate code to calculate the Fibonacci sequence to 100." "Execute it and give me the result." }] response = client.messages.create( max_tokens=1024, model="claude-3-5-sonnet-20240620", messages=messages, # Provides a sandboxed code interpreter tools=th.get_tools(), ) messages += th.run_tools(response)# OpenAI Completions th = Toolhouse() client = OpenAI() messages = [{ "role": "user", "content": "Look at my Linear backlog and send a standup update on Slack." }] response = client.chat.completions.create( max_tokens=1024, model="gpt-4o", messages=messages, # Provides Linear and Slack API integrations tools=th.get_tools(), ) messages += th.run_tools(response)# Llama on Groq th = Toolhouse() client = Groq() messages = [{ "role": "user", "content": "Get the temperature record for San Francisco," "compute the median, and give me the result." }] response = client.chat.completions.create( model="llama3-70b-8192", messages=messages, # Provides weather API and a sandboxed code interpreter tools=th.get_tools(), ) messages += th.run_tools(response)# Command-r th = Toolhouse() client = cohere.Client() message = "I'm hungry! Can you give me a recipe for one of my favorite food?" response = client.chat( model="command-r", message=message, # Provides a memory tool tools=th.get_tools(), ) tool_results = th.run_tools(response)# Anthropic th = Toolhouse() client = Anthropic() messages = [{ "role": "user", "content": "Generate code to calculate the Fibonacci sequence to 100." "Execute it and give me the result." }] response = client.messages.create( max_tokens=1024, model="claude-3-5-sonnet-20240620", messages=messages, # Provides a sandboxed code interpreter tools=th.get_tools(), ) messages += th.run_tools(response)# OpenAI Completions th = Toolhouse() client = OpenAI() messages = [{ "role": "user", "content": "Look at my Linear backlog and send a standup update on Slack." }] response = client.chat.completions.create( max_tokens=1024, model="gpt-4o", messages=messages, # Provides Linear and Slack API integrations tools=th.get_tools(), ) messages += th.run_tools(response)# Llama on Groq th = Toolhouse() client = Groq() messages = [{ "role": "user", "content": "Get the temperature record for San Francisco," "compute the median, and give me the result." }] response = client.chat.completions.create( model="llama3-70b-8192", messages=messages, # Provides weather API and a sandboxed code interpreter tools=th.get_tools(), ) messages += th.run_tools(response)# Command-r th = Toolhouse() client = cohere.Client() message = "I'm hungry! Can you give me a recipe for one of my favorite food?" response = client.chat( model="command-r", message=message, # Provides a memory tool tools=th.get_tools(), ) tool_results = th.run_tools(response)# Anthropic th = Toolhouse() client = Anthropic() messages = [{ "role": "user", "content": "Generate code to calculate the Fibonacci sequence to 100." "Execute it and give me the result." }] response = client.messages.create( max_tokens=1024, model="claude-3-5-sonnet-20240620", messages=messages, # Provides a sandboxed code interpreter tools=th.get_tools(), ) messages += th.run_tools(response)# OpenAI Completions th = Toolhouse() client = OpenAI() messages = [{ "role": "user", "content": "Look at my Linear backlog and send a standup update on Slack." }] response = client.chat.completions.create( max_tokens=1024, model="gpt-4o", messages=messages, # Provides Linear and Slack API integrations tools=th.get_tools(), ) messages += th.run_tools(response)# Llama on Groq th = Toolhouse() client = Groq() messages = [{ "role": "user", "content": "Get the temperature record for San Francisco," "compute the median, and give me the result." }] response = client.chat.completions.create( model="llama3-70b-8192", messages=messages, # Provides weather API and a sandboxed code interpreter tools=th.get_tools(), ) messages += th.run_tools(response)# Command-r th = Toolhouse() client = cohere.Client() message = "I'm hungry! Can you give me a recipe for one of my favorite food?" response = client.chat( model="command-r", message=message, # Provides a memory tool tools=th.get_tools(), ) tool_results = th.run_tools(response)
Testimonials
Testimonials
Testimonials
Why developers love Toolhouse
Defining functions is annoying. You'll need to understand the prompt for the function, and build the payload. Toolhouse does that for you.
Yoshimasa Niwa
Software Engineer, X
I like the simplicity of the implementation! We can focus more on building and improving features rather than managing configuration and integration.
Jacky C
Software Developer
This makes so much sense! LLM providers diverge their implementation for no reason, and Toolhouse brings sanity to my codebase.
CJ Avilla
Software Engineer, Craftwork
I love the marketplace — it's like having npm, but is also shields the complexity away from developers. It helps any models that supports function calling be more powerful!
Jamie Maguire
Engineer, Microsoft MVP, AI
As a software development agency, I like Toolhouse because it allows me to build tools and functions for my client without giving them IP on my own tools.
Fabio Pulvirenti
CEO, Etiqa
I love that you give the utmost flexibility in terms of how the data is handled. Toolhouse works across LLMs and that's a fantastic reassurance in terms of privacy.
Jim Bennett
Head of Developer Advocacy, Pieces
Having tried building a few tools myself, I know how helpful those prebuilt components can be. LangChain doesn't do that. GPTs can't do that.
Zach Hamed
Founder & Co-CEO, Clay
Defining functions is annoying. You'll need to understand the prompt for the function, and build the payload. Toolhouse does that for you.
Yoshimasa Niwa
Software Engineer, X
I like the simplicity of the implementation! We can focus more on building and improving features rather than managing configuration and integration.
Jacky C
Software Developer
This makes so much sense! LLM providers diverge their implementation for no reason, and Toolhouse brings sanity to my codebase.
CJ Avilla
Software Engineer, Craftwork
I love the marketplace — it's like having npm, but is also shields the complexity away from developers. It helps any models that supports function calling be more powerful!
Jamie Maguire
Engineer, Microsoft MVP, AI
As a software development agency, I like Toolhouse because it allows me to build tools and functions for my client without giving them IP on my own tools.
Fabio Pulvirenti
CEO, Etiqa
I love that you give the utmost flexibility in terms of how the data is handled. Toolhouse works across LLMs and that's a fantastic reassurance in terms of privacy.
Jim Bennett
Head of Developer Advocacy, Pieces
Having tried building a few tools myself, I know how helpful those prebuilt components can be. LangChain doesn't do that. GPTs can't do that.
Zach Hamed
Founder & Co-CEO, Clay
Introducing the platform
Introducing the platform
Introducing the platform
What is Toolhouse?
Introducing the best developer experience for implementing function calling across any LLM.
A marketplace of tools
Easily integrate function calls for LLMs. Find the tools you need as a developer to supercharge your application. Publish and monetize your existing tools.
A marketplace of tools
Easily integrate function calls for LLMs. Find the tools you need as a developer to supercharge your application. Publish and monetize your existing tools.
A marketplace of tools
Easily integrate function calls for LLMs. Find the tools you need as a developer to supercharge your application. Publish and monetize your existing tools.
A hosting platform
We host tools that you can rely on: whether you need to scrape data from the web, implement long-term memory or send emails.
A hosting platform
We host tools that you can rely on: whether you need to scrape data from the web, implement long-term memory or send emails.
A hosting platform
We host tools that you can rely on: whether you need to scrape data from the web, implement long-term memory or send emails.
An SDK for developers
Save heaps of time when implementing function calling. We support all major LLMs. Stop rewriting boilerplate code every time you switch model.
An SDK for developers
Save heaps of time when implementing function calling. We support all major LLMs. Stop rewriting boilerplate code every time you switch model.
An SDK for developers
Save heaps of time when implementing function calling. We support all major LLMs. Stop rewriting boilerplate code every time you switch model.
The Tool Store
The Tool Store
The Tool Store
LLM tools done for you
Speed up time to market by unleashing developer productivity, automating and streamlining AI function calling. Works with any LLM that supports function calling.
✓ Tool discovery: A curated marketplace of useful tools to supercharge your AI code.
✓ Tool evaluation: Find the best tool for any job. Access analytics and documentation instantly.
✓ Tool integration: Integrate one or more tools in seconds, with the power of our SDK and Tool Bundles.
Universal SDK
Universal SDK
Universal SDK
Integrate tools quickly
Toolhouse brings the most delightful developer experience in AI.
✓ Implement tools fast: We built our SDKs with the best developer experience in mind.
✓ One SDK, many LLMs: Streamline LLM function calls management across multiple models.
✓ Built-in analytics: Simplify debugging with comprehensive logging of tool inputs and outputs.
Developer Tools
Developer Tools
Developer Tools
A platform that scales
Our optimized infrastructure makes it easy to improve your overall LLM performance.
✓ Define and optimize tools once, use them with any supported LLM.
✓ Stability matters: Our tools are hosted on reliable and scalable infrastructure, and are constantly tested against any supported LLM.
✓ Cost effective and transparent: Your usage can easily scale based on your volume. We support caching to help you save money and decrease latency.
Benefits
Benefits
Benefits
Expand the capabilities of your LLM
Toolhouse takes care of integrating the best tools to your LLMs, so you don't have to. Focus on building the best experiences for your users. Toolhouse takes care of the rest.
Faster development cycles
Boost development speed and slash costs with our intuitive platform. Cut weeks off your timeline, prototype rapidly, and seamlessly integrate function calling in your code.
Faster development cycles
Boost development speed and slash costs with our intuitive platform. Cut weeks off your timeline, prototype rapidly, and seamlessly integrate function calling in your code.
Faster development cycles
Boost development speed and slash costs with our intuitive platform. Cut weeks off your timeline, prototype rapidly, and seamlessly integrate function calling in your code.
Tool Exploration
Access a curated marketplace of high-quality, time-saving tools. Easily integrate and manage tools using our SDK.
Tool Exploration
Access a curated marketplace of high-quality, time-saving tools. Easily integrate and manage tools using our SDK.
Tool Exploration
Access a curated marketplace of high-quality, time-saving tools. Easily integrate and manage tools using our SDK.
Debugging and Observability
Streamline debugging and traceability for your AI systems. Resolve issues quickly.
Debugging and Observability
Streamline debugging and traceability for your AI systems. Resolve issues quickly.
Debugging and Observability
Streamline debugging and traceability for your AI systems. Resolve issues quickly.
Reliable Infrastructure
Leverage our high-performance infrastructure for seamless operations. Enjoy robust scalability for low-latency responses generation
Reliable Infrastructure
Leverage our high-performance infrastructure for seamless operations. Enjoy robust scalability for low-latency responses generation
Reliable Infrastructure
Leverage our high-performance infrastructure for seamless operations. Enjoy robust scalability for low-latency responses generation
Control and Analytics
Enjoy transparent function calling across multiple LLMs. Leverage tools for trustworthy code execution or to reduced AI hallucinations with memory.
Control and Analytics
Enjoy transparent function calling across multiple LLMs. Leverage tools for trustworthy code execution or to reduced AI hallucinations with memory.
Control and Analytics
Enjoy transparent function calling across multiple LLMs. Leverage tools for trustworthy code execution or to reduced AI hallucinations with memory.
Developer friendly pricing
A free tier to get started. A friendly pricing experience that scales with usage. Just like developers love.
Developer friendly pricing
A free tier to get started. A friendly pricing experience that scales with usage. Just like developers love.
Developer friendly pricing
A free tier to get started. A friendly pricing experience that scales with usage. Just like developers love.
Frequently asked questions.
Most commonly asked questions about the Toolhouse experience.
Most commonly asked questions about the Toolhouse experience.
Most commonly asked questions about the Toolhouse experience.
What languages does the Toolhouse SDK support?
We're starting with Python, and we will release more SDKs and frameworks soon.
What are the supports i get?
We're starting with Python, and we will release more SDKs and frameworks soon.
What's the pricing model?
Are tools on the Tool Store safe?
Is my data kept private?
Get early access today
Sign up today and join a better way to build smarter agents.
Get early access for $1 and receive $150 in Toolhouse execution credits.
Payment is optional to access rewards
Fully refundable, no questions asked
Product
Info
© 2024 Toolhouse Technologies, Inc. All rights reserved.
Product
Info
© 2024 Toolhouse Technologies, Inc. All rights reserved.
Product
Info
© 2024 Toolhouse Technologies, Inc. All rights reserved.