Features

• No model call, no tokens. The import runs in a Claude Code hook and prints its report directly. Tokens are only spent once you resume a session and start typing, just like any other conversation.

• No duplicates, ever. A session whose content hasn’t changed since its last import is skipped. A session that changed gets a new copy with the next suffix: Title, Title (1), Title (2)…. Existing files are never overwritten.

• Only this repository. Scoping works like /resume: the current git worktree by default, or --worktrees for all worktrees of the repo.

• Reads opencode’s SQLite database read-only: opencode 1.2+ (1.x and 2.x layouts), with opencode export as a fallback.

• No dependencies, no network access.

Prerequisites

Python 3.10+

Installation

claude plugin marketplace add barseghyanartur/claude-opencode-sessions
claude plugin install opencode-sessions@barseghyanartur

/plugin marketplace add barseghyanartur/claude-opencode-sessions
/plugin install opencode-sessions@barseghyanartur

uv tool install claude-opencode-sessions      # or: pipx install claude-opencode-sessions
uvx claude-opencode-sessions list             # one-off run without installing

Usage in Claude Code

/opencode-sessions:import                      # sessions of the current worktree
/opencode-sessions:import --worktrees          # every worktree of the repository
/opencode-sessions:import --dry-run            # show what would happen
/opencode-sessions:import --with-tool-output   # include (truncated) tool output
/resume                                        # pick an "[opencode] …" session

Example report:

opencode → Claude Code import · git worktree /Users/me/repos/brrn
into ~/.claude/projects/-Users-me-repos-brrn

  imported     [opencode] Fixing stalled make armdict-fetch-hy build (1)
  unchanged    [opencode] HTTP Request/Response Log Analysis for brrn.ru …

1 imported, 1 unchanged · skipped: 1 sub-agent
Open them with /resume — titles start with "[opencode]".

Command line

The same package works in the terminal:

claude-opencode-sessions import --dry-run           # same as /opencode-sessions:import
claude-opencode-sessions import --worktrees
claude-opencode-sessions list                       # sessions in scope, newest first
claude-opencode-sessions list --all --format json
claude-opencode-sessions doctor                     # environment + database report

Hidden by default: archived sessions (--include-archived) and sub-agent sessions (--include-subagents).

How it works

• sqlite (default): opens opencode.db with mode=ro. On 2.x it reads session_v2 + session_message. On 1.x it reads session + message + part. When both exist in the same file, sessions are merged by id (2.x wins), and messages fall back to the 1.x tables per session.

• cli (fallback): used when the database is missing, locked or not recognised, or when more than 20% of a session’s messages can’t be parsed. It runs opencode session list --format json and opencode export <id>.

See docs/design.md for the design and docs/schema-notes.md for the tables and fields that are read.

Development

make install     # uv sync (creates .venv with dev tools)
make test        # pytest (tests live in src/claude_opencode_sessions/tests)
make check       # ruff + mypy --strict + pytest + manifest validation
make run ARGS="list --all"
make dev         # claude --plugin-dir . (live plugin; /reload-plugins after edits)
make install-local / make uninstall-local

Troubleshooting

If /opencode-sessions:import makes Claude answer instead of printing a report, the hook didn’t run. Check the Python requirement above and run claude-opencode-sessions doctor, or sh ~/.claude/plugins/…/scripts/opencode-sessions doctor. It shows the Python in use, the database path, the detected layout, row counts and the scope roots for the current directory.

License

MIT

Support

For security issues contact me at the e-mail given in the Author section.

For overall issues, go to GitHub.

Author

Artur Barseghyan <artur.barseghyan@gmail.com>