Poolside API
Product · API / developer platform by Poolside
Bring frontier agentic coding models to your applications.
An OpenAI-compatible API providing programmatic access to Poolside's Laguna agentic coding models. It enables developers to integrate long-horizon reasoning and autonomous coding capabilities into custom applications and tools.
- Price
- From 0 USD
- Languages
- English
Category: AI Infrastructure
- APIs and SDKs
- OpenAI-compatible Chat API
- Platforms
- API
- Deployment
- Cloud (inference.poolside.ai)
- Languages
- Python, TypeScript, Go, Rust, Java
- Base URL
- https://inference.poolside.ai/v1
- Compatibility
- OpenAI-compatible Chat API
- OpenAI-compatible Chat API for seamless integration
- Access to Laguna S 2.1 with 1M context window
- Access to Laguna XS 2.1 for fast agentic iteration
- Free to use for a limited time during research preview
Overview
The Poolside API provides developers with programmatic access to Poolside's frontier agentic coding models, including the Laguna series. It features an OpenAI-compatible chat completion endpoint, allowing for integration into existing workflows and tools that already support the OpenAI specification. Developers can use the API to power custom applications, automated scripts, and internal developer tools with long-horizon reasoning capabilities. The API is hosted at inference.poolside.ai and serves as a primary interface for building third-party applications on top of Poolside's foundation models. It is currently available in research preview, offering access to models like Laguna S 2.1 and Laguna XS 2.1 for software engineering tasks.
APIs and Model Access
- OpenAI Compatibility: The API is fully compatible with OpenAI's client libraries, allowing developers to use existing OpenAI SDKs by changing the base URL.
- Laguna S 2.1 Access: Developers can access the Laguna S 2.1 model for long-horizon agentic coding and complex software engineering tasks.
- Laguna XS 2.1 Access: The API provides access to Laguna XS 2.1, a lightweight and fast model optimized for rapid agentic iteration.
- Inference Endpoint: The API endpoint is hosted at inference.poolside.ai/v1 for programmatic model requests.
Integration and Development
- Chat Completions: The API supports chat completion requests for generating code, explaining concepts, and solving software problems.
- Custom Tooling: Developers can integrate Poolside's intelligence into custom applications, automated scripts, and internal developer tools.
- Agentic Workflows: The API enables the creation of agentic workflows that can perform multi-step reasoning and autonomous task execution.
- Developer Experience Integration: The API can be used to power coding assistants in various environments, including IDEs and terminal-based tools.
Use cases
- Developers integrate the API into custom IDE plugins to provide context-aware code completions and refactoring.
- Engineering teams use the API to automate the generation of unit tests and documentation for large codebases.
- Software architects leverage the 1M context window of Laguna S 2.1 to analyze and modernize legacy systems.
- DevOps engineers build automated scripts that use the API to audit code for race conditions and security flaws.
Specifications
- APIs and SDKs
- OpenAI-compatible Chat API
- Platforms
- API
- Deployment
- Cloud (inference.poolside.ai)
- Languages
- Python, TypeScript, Go, Rust, Java
- Base URL
- https://inference.poolside.ai/v1
- Compatibility
- OpenAI-compatible Chat API
- Supported Models
- Laguna S 2.1, Laguna XS 2.1
- Laguna S 2.1 Context Window
- 1,000,000 tokens
- Laguna XS 2.1 Context Window
- 256,000 tokens
- Authentication
- API Key (POOLSIDE_API_KEY)
How it works
- Get an API Key: Obtain a POOLSIDE_API_KEY from the Poolside platform to authenticate your requests.
- Configure the Endpoint: Configure your application or the OpenAI SDK to use https://inference.poolside.ai/v1 as the base URL.
- Call the API: Send chat completion requests specifying models like poolside/laguna-s-2.1 to receive agentic coding responses.
Pricing
- Research Preview
- Free to use for a limited time
FAQ
How do I integrate the Poolside API into my existing OpenAI-based project?
The Poolside API is fully compatible with OpenAI's client libraries. Developers can use the standard OpenAI SDK and simply point the base_url to https://inference.poolside.ai/v1 while providing their POOLSIDE_API_KEY.
Which models are available through the Poolside API?
The API currently provides access to the Laguna family of models, including Laguna S 2.1 (118B parameters) and Laguna XS 2.1 (33B parameters).
What is the cost of using the Poolside API?
The Poolside API is currently free to use for a limited time as part of a research preview.
What programming languages can the models handle via the API?
The API supports a variety of programming languages for coding tasks, including Python, Rust, Java, and Go, which are part of the models' training corpus.
Added by AntSpace from public sources (poolside.ai).