Kiln

v1.0.2 Feature

This release adds 2 notable features for engineering teams evaluating rollout.

Published 23d AI Coding Tools

View tool

✓ No known CVEs patched

Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

ai chain-of-thought collaboration dataset-generation evals evaluation

+13 more

evaluation-framework fine-tuning machine-learning macos mcp ollama openai prompt prompt-engineering python rlhf synthetic-data windows

ReleasePort's take

Light signal

editorial:auto 13d

Kiln v1.0.2 adds Synthetic Data Guides for structured synthetic data generation. Integrates GPT-5.5, DeepSeek V4 Pro/Flash, and Gemini Embedding 2 models.

Why it matters: Expand model choice across providers and enable richer synthetic data generation. Evaluate new models in dev environments; integrate Gemini Embedding 2 if you need current embedding alternatives.

Summary

AI summary

Synthetic data guides, new models GPT-5.5, DeepSeek V4 Pro and Flash, Gemini Embedding 2 added.

Changes in this release

Type	Severity	Summary	CVE
Feature
Feature	Medium	GPT-5.5 model now supported GPT-5.5 model now supported Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	DeepSeek V4 Pro model now supported DeepSeek V4 Pro model now supported Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	DeepSeek V4 Flash model now supported DeepSeek V4 Flash model now supported Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	Gemini Embedding 2 model now supported Gemini Embedding 2 model now supported Source: llm_adapter@2026-05-21 Confidence: high	—
Feature	Medium	Synthetic Data Guides help generate more complex synthetic data with structure, tone, guidelines, rules. Synthetic Data Guides help generate more complex synthetic data with structure, tone, guidelines, rules. Source: llm_adapter@2026-05-21 Confidence: low	—

Full changelog

Kiln V1.0 🎉

After 6,562 commits — Kiln is hitting v1.0!

Unlike most releases, V1 isn't about a pile of new features. It's a mark of stability and robustness. APIs have been battle tested, core flows verified.

What is Kiln?

Kiln is the AI workbench for teams — everything you need to build and optimize AI systems, in one app and one open-source library:

Build — RAG, Tools & MCP, Skills, Sub-Agents, structured output, reasoning, prompt generators
Evaluate — LLM-as-Judge evals, AI Eval Builder, golden datasets, human ratings
Improve — Auto-Optimize prompts against your evals, fine-tuning, synthetic data generation
Collaborate — Git Auto-Sync, in-app feedback & ratings, MIT-licensed Python library
Run anywhere — 190+ models tested across every major provider, cloud or fully local

Friendly app for the whole team (PMs, SMEs, data scientists). MIT-licensed Python library (3.10+) for production. Same project files, no lock-in.

Thanks

To every contributor, every issue reporter, every Discord member, and everyone who starred the repo — this release belongs to you. Onwards.

What's New in V1.0

Synthetic Data Guides: Guides help you generate more complex synthetic data. Kiln will help you build a data-guide covering structure, tone, guidelines, and rules.
New Models: GPT-5.5 , DeepSeek V4 Pro and Flash, Gemini Embedding 2
And More: visual design updates, UX polish, bug fixes, and more

CI Build Source for this release Mac and Linux: /Kiln-AI/Kiln/actions/runs/25699347547
CI Build Source for this release Windows: /Kiln-AI/Kiln/actions/runs/25699347566

Full Changelog: https://github.com/Kiln-AI/Kiln/compare/v0.28.0...v1.0.2

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track Kiln

Get notified when new releases ship.

About Kiln

All releases →