SageRift Launch Roadmap: 2026 & Beyond

Making coding agents safe to run

Today marks a pivotal moment for SageRift. After months of heads-down development and a tight private alpha, we're ready to share where SageForge is headed.

Our mission is simple but ambitious: let you run AI coding agents on real repos without the mess. Every agent gets an isolated worktree, a scoped set of files it may touch, and a review gate that won't let unverified work land. That safety core — and the workflow around it — is what we're building in the open.

Below is our detailed roadmap for the year. We believe in building in public, which means sharing not just the wins, but the plans, the delays, and the trade-offs.

---

---

What to expect in Q1

The first quarter is focused on the foundation: the worktree + scope + review-gate core that makes agents safe to run. For our alpha testers, that means SageForge — scope a task, let Claude Code, Codex, or pi work in an isolated worktree, run your checks, and approve a clean diff.

Key focus areas

1. The safety core: One worktree per task, diff-time file-scope enforcement, and a backend-enforced review gate.
2. Local-first by default: No cloud, no telemetry. Your repos and history stay on your machine; you own every Git action.
3. Documentation: Great tools deserve great docs. We're investing heavily in our Docs and Learn sections.

A note on early access

During Q1, access stays invite-only while we harden the Windows alpha. We're prioritizing reliability over rapid growth. Pro early-access members get builds first.

If you want in, join the waitlist — and if you're rolling SageForge out across a team, talk to us about priority onboarding.

Join the discussion

We're building in public. Join the community on Discord or follow development on GitHub. Your feedback shapes the roadmap.

Happy building.

The SageRift Team