
Tessera
Command Center for AI Coding Agents
Hi Indie Hackers,
I’m building Tessera, an open-source web and desktop workspace for CLI-based coding agents like Claude Code, Codex, and OpenCode.
I started building it because TUI-based agent workflows did not really fit how I work. The agents were useful, but the work around them became hard to manage once I had several tasks, branches, tool calls, and file changes happening at the same time.
In a terminal, coding agent work often feels temporary. You start something, switch context, come back later, and then have to remember what the agent was doing, which files changed, what failed, and whether the work is ready to review.
Tessera turns that work into something you can organize and come back to.
It lets you:
- keep agent work visible in list and Kanban views
- group related work into collections
- open agents, terminals, and files in tabs and split panes
- connect tasks to Git worktrees for isolation
- inspect tool logs, failures, diffs, branches, and PR state
- use Claude Code, Codex, and OpenCode from one workspace
The goal is not to replace coding agents. It is to make the work around them easier to organize, resume, and review.
I’ve been using Tessera for product research, development, and day-to-day company work. I open-sourced it because it became too useful to keep as a personal tool.
GitHub:
https://github.com/horang-labs/tessera
I’d love feedback from anyone using coding agents in real projects, especially around worktree workflows and what state a GUI should preserve around agent work.
About
I built Tessera because CLI coding agents are powerful, but their sessions become hard to manage once real work starts. Tessera keeps agent sessions, tasks, worktrees, diffs, and PRs organized.

Comment