Ridges (SN62) Ridgeline Ships Four New Developer Features
Ridges (SN62) Ridgeline just dropped a significant product update. The AI coding agent platform built on Bittensor Subnet 62 now includes four new features designed to give developers tighter control over their agent runs. The update brings Rerun, Stop, a collapsible chat view, and per-run commit stats to the platform.
For those catching up, Ridgeline is the product layer that takes competitive AI coding agents from Ridges (SN62) and puts them to work for real developers. The platform integrates directly with GitHub, allowing developers to assign issues to AI agents that autonomously read repositories, generate patches, and test code. We covered the initial launch in detail in our earlier piece, Ridgeline: Ridges (SN62) Turns AI Agents Into a Developer Tool.
What Is New in This Update
The four features all target the same problem: giving developers more granular control over how agent runs behave. Here is what each one does.
Rerun lets developers add more context to a previous run and execute it again. This is useful when a response needs updating or when new requirements emerge mid-task. Instead of starting from scratch, a developer can layer in additional instructions and let the agent iterate on its previous work.
Stop allows developers to kill a running agent at any point. If something comes to mind mid-run, there is no need to wait for it to finish. A developer can stop the run, add new context, and rerun immediately. The team frames this as a way to eliminate wasted compute.
Collapsible chat view organizes all runs tied to a single issue into a clean, threaded interface. Every run for a given GitHub issue now appears in a chat-like format. This makes it easy to review the full history of an issue without clutter.
Commit stats per run adds detailed commit data to each individual run. Developers can now see exactly what changed, with diffs opening in a new tab for quick review.
Why This Matters
These additions mark a clear shift in the Ridges (SN62) Ridgeline experience. The initial open beta introduced the core loop of assigning GitHub issues to AI agents and letting them work autonomously. This update now adds a layer of real-time developer oversight to that loop.
The Rerun and Stop features together create a feedback cycle that mirrors how developers already work with human teammates. Assign a task, review the output, course-correct if needed, and iterate. The collapsible chat view and commit stats round this out by making the history of each task transparent and easy to navigate.
From a broader Bittensor ecosystem perspective, Ridgeline continues to demonstrate what a front-door application built on subnet infrastructure can look like. The platform still runs inference through Targon (SN4) and Chutes (SN64), composing across multiple subnets to deliver its core functionality. Each product update like this one brings the experience closer to what mainstream developer tools already offer, which is exactly the kind of traction the ecosystem needs.
The Ridgeline open beta remains free to use, with 10 free credits for every new account. Try it here.
To learn more about the subnet behind the product, including how miners compete and how the incentive mechanism works, check out our Simple Guide to Ridges (SN62).
