Introducing Seahaven: Synthetic Worlds for Agent Evals and RL
Seahaven is an open source Python framework for building synthetic worlds for AI agents: stateful, resettable, parallel environments for evals and RL.
Kiln Blog5 min read
Kiln
29 stories
Auto-optimizing Jev: half the errors, 1/7 the cost
We pointed our autoresearch loop at Jev, a classifier that can't be fine-tuned. It ended up making half as many mistakes as a reasoning model, for a seventh of the cost and a thirtieth of the latency.
Kiln Blog17 min readWe made our launch video with Claude Code (and open-sourced the tool)
Kiln just launched v1. We made the launch video in Claude Code. Along the way we built and open-sourced videowright: the tool we wished we'd had.
Kiln Blog6 min readGit-Backed SaaS: SaaS Without the Service
Kiln uses git for everything you'd normally need a database for, with the instant-syncing user-experience of SaaS. Every API call is a commit. Every read is a local file.
Kiln Blog11 min readWhy Your Agent Probably Needs Skills
Kiln has added agent skills support! Here's what skills are, why progressive disclosure works, and how to use them in your agents.
Kiln Blog3 min readIntroducing Kiln Prompt Optimizer
Kiln will now automatically find the ideal prompt for your task, guided by your own evals. Performance gains that blow away manual optimization, and even beat fine-tuning.
Kiln Blog3 min readThe Requirements Layer Your AI System is Missing
Most AI teams think they have a testing problem. But the real issue is they aren't defining what "good" actually means. Learn how specs can transform your AI development workflow.
Kiln Blog7 min readKiln Eval Builder & AI Assistant for Evals and Synthetic Data
Introducing Kiln Eval Builder — a guided assistant that helps you build accurate evals, synthetic data, and judge prompts in minutes instead of hours.
Kiln Blog2 min readRAG Isn't One Size Fits All: Here's How to Tune It for Your Use Case
Great RAG comes from a tight iteration loop. Learn how to systematically improve each layer of your RAG system using Kiln and LanceDB.
Kiln Blog13 min readNew in Kiln: RAG Q&A Evals, Tool Use Evals, Reranking, Semantic Chunking, an MCP server, and more!
Our new v0.23.0 app release includes powerful new features for RAG, evals, and more!
Kiln Blog3 min readIntroducing Kiln Agent Builder
Kiln now supports building agentic systems in as little as 5 minutes. Also new is Local RAG: build RAG systems without any third party services.
Kiln Blog3 min readNew in Kiln: Search Tools/RAG, Document Library, and Synthetic Data V3
Introducing our biggest release ever! Build state-of-the-art RAG systems in minutes, with drag-and-drop!
Kiln Blog2 min readNew in Kiln: Tool Calling and MCP
We just added support for tool-calling and MCP (Model Context Protocol) to Kiln! This new feature includes:
Kiln Blog2 min readEnd to End AI Project with Evals, Synthetic Data, and Fine-Tuning
A video walk through of Kiln project from start to finish. The project we build is open source and available on GitHub.
Kiln Blog8 min readI wrote 2000 LLM test cases so you don't have to
I've been building Kiln AI: an open tool to help you find the best way to run your AI workload. Testing different models to see which ones work best is a huge part of this process.
Kiln Blog7 min readKiln Release v0.18.1: Evals V2, Synthetic Data V2, Kiln Issues, New Models, and more!
Our new release is our biggest ever! Read about how we improved evals, synthetic data, and a whole new concept: Kiln Issues.
Kiln Blog3 min readMany Small Evals Beat One Big Eval, Every Time
Evaluating AI products is notoriously difficult. In this article we walk through how to make it approachable, maintainable, and iterative using many small evals instead of one large eval.
Kiln Blog8 min readDistillation in Practice: Ablating Gemma 3 27B with synthetic data from Sonnet 4
A hands-on experiment fine-tuning Gemma 3 27B using synthetic data from Sonnet 4, achieving performance that rivals GPT-4o on specific tasks while being smaller and cheaper to run.
Kiln Blog5 min read
Showing 18 of 29 stories
Load more storiesAiByts Weekly Brief
Get the top AI news and tools straight to your inbox every week.
Weekly · AI news
Join 15,000+ AI enthusiasts