Letter Models Lets Companies Fine-Tune on Their Voice in Three Clicks
Scobleizer · x · 2026-11-17
Letter released Letter Models, pitched as the fastest way to train a model on a company's voice. The premise: coding agents are good at building but bad at writing.
The workflow is three steps — ingest examples of company voice (2 min), pick a base model (1 min), choose use cases (1 min). It ships with Qwen 3.5 27B and Gemma 3 27B, creates a LoRA adapter for each, and benchmarks them against Claude / GPT with skills. Output can be used immediately with coding agents like Claude Code or Codex under segmented duties; setup takes under five minutes.
More from coding & agent
- Every staffer clones her editors from Google Docs edit history in a word processor built on Fable 5.2 — danshipper · 2026-11-17
- YC-backed OpenCrawl launches open source, letting coding agents search and talk to each other — ycombinator · 2026-11-17
- Stealth models roadrunner and silicon-2 spotted in Devin Desktop build, makers unconfirmed — lyraxana · 2026-11-17
- Personal AI agents are so proactive that one email triggers 10 notifications, users report — altryne · 2026-11-17
- Every's Frontier Team Logs Weekly Agent Experiments, From Map-Driven Research to Screen-Recording Handoff — kieranklaassen · 2026-11-17
- WebArena-Pro debuts as largest browser-agent benchmark, top agents score 62.3% — xhluca · 2026-11-17