@julien_c: Today I'm launching a new project called SynthTraces It is a minimal codebase to generate synthetic coding agent sessio…
Summary
Julien Chaumond launches SynthTraces, a minimal codebase that generates synthetic coding agent session traces by having an open model (via HF Inference Providers) interact with a small local model (via llama.cpp) on real open-source codebases, producing over 2,000 Pi session traces for training and fine-tuning LLMs.
View Cached Full Text
Cached at: 06/05/26, 07:11 AM
Today I’m launching a new project called SynthTraces
It is a minimal codebase to generate synthetic coding agent session traces using Pi (from @badlogicgames)
I wanted a large number of coding-agent traces, so I built a tiny harness where two models talk to each other:
- an open model (served via HF Inference Providers) plays the coding agent. It gets read + bash access to a real open source codebase (the huggingface OSS projects)
- a small local model (llama.cpp) plays the human user, asking simple questions like “how do I run this?” or “how is CI set up?”
The result is more than 2,000 Pi session traces which can be used to train or fine-tune LLMs, and optimize them for Pi
And ofc everything is published on @huggingface
Similar Articles
@benhylak: we built the first sane way to debug your agent locally. you can see your traces. codex/claude code can too. this lets …
A new open source tool enables local debugging of AI agents by viewing traces, allowing automated eval writing and testing with tools like codex and Claude code.
@ClementDelangue: We need open traces so that everyone can train open agent models! cc @steipete @badlogicgames @thdxr @matanSF @hwchase17
Clement Delangue advocates for open traces to democratize training of open agent models.
I built a tool to turn your Claude Code sessions into fine-tuning data for local models
A new open-source tool called claude_converter converts Claude Code session logs into fine-tuning datasets compatible with TRL/SFTTrainer, Axolotl, and LLaMA-Factory, enabling developers to repurpose real coding conversations for training local models.
I rebuilt a Claude Code–style coding agent from scratch — the whole agent loop is 6 lines. 20 chapters, ~5k lines, no frameworks, runs on local models too
A developer shares a 20-chapter tutorial rebuilding a Claude Code–style coding agent from scratch, showing the entire agent loop in ~6 lines, with support for local models and multiple LLM APIs.
@EBorgnia: Today we're launching Jacq. A coding agent built together with the small models we've been training at @relace_ai for t…
Jacq is a cloud-based coding agent that integrates with Slack, Linear, GitHub, email, and other tools, using small models trained by Relace AI to pull context from connected devices and maintain durable threads for work history.