DATABASES Signal 498
Show HN: Wallfacer – A terminal session manager for Claude Code, and more
Wallfacer adds a searchable, tag-gable index for AI coding session files from Claude Code, Cursor CLI, Kiro CLI and Codex.
Developers who run AI-assisted coding tools accumulate many JSONL transcripts that are hard to differentiate. Wallfacer centralises those files in a read-only SQLite store, letting users locate, rename, tag, and resume sessions from a terminal UI or scriptable CLI. Because it never mutates the original agent files, it can be adopted with minimal risk, but it only works for the four supported agents and on macOS/Linux environments.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
Wallfacer builds a read-only SQLite index of session files from four AI coding agents, preserving original files untouched.
It offers both an interactive full-screen TUI and a set of CLI subcommands for creating, searching, tagging, resuming, and safely deleting sessions.
Installation is via a Homebrew tap or a Go install, requiring Go 1.22+ for builds and running on macOS or Linux.
THE READ
What the cluster adds up to.
AI coding assistants such as Claude Code, Cursor CLI, Kiro CLI and Codex each dump conversation logs into their own directory structures, resulting in a scattered collection of JSONL files. Wallfacer scans those locations, builds a read-only SQLite index, and stores user-supplied metadata like titles, tags, and project names. This central catalogue lets engineers see every session across agents in a single chronological view without altering the original files.
The tool presents two front-ends that share the same index: a full-screen terminal UI for interactive browsing and a set of one-shot CLI commands for automation. Users can start new sessions, resume existing ones, filter by agent, tag, or project, and perform safe deletions that move files to a trash area unless a purge flag is used. Because the UI and CLI operate on the same database, any metadata added in one mode is instantly available to the other.
Wallfacer can be installed through a dedicated Homebrew tap or by invoking Go’s install command, with pre-built binaries also provided. The only runtime requirement beyond the binary is a recent Go toolchain for building from source, and the software runs on macOS and Linux platforms. Adoption cost is limited to the installation step and the disk space for the SQLite index, which is modest compared to the size of the underlying session logs.
Support is currently limited to the four named agents, and the indexing logic depends on their specific file-layout conventions. Wallfacer does not modify the agents’ transcript files, so it cannot recover sessions that are stored outside the expected directories or in formats it does not recognise. The tool also does not claim Windows compatibility, so users on that OS would need an alternative solution.
For engineers who regularly invoke AI-driven coding sessions, Wallfacer provides a practical way to organise and retrieve past work, reducing time spent hunting for the right transcript. The CLI subcommands enable integration into scripts or CI pipelines, while the TUI offers a quick visual overview. Overall, the barrier to entry is low, but extending coverage to additional agents will require writing new adapters.
Written by elseif from the cluster below · checked for specifics the sources never containedTHE CLUSTER