Search the local gadak mirror of Jira and Confluence from Raycast and open a hit in the Gadak app.
Each keystroke runs gadak search --json --limit 20 against the on-disk
mirror. Results come in two sections: issues, then documents (Confluence
pages). Enter opens gadak://view?issue=<KEY> for an issue and
gadak://view?doc=<ID> for a document. When a profile preference is set,
the links gain a /w/<profile> segment.
Extensions like Jira Search talk to the Atlassian REST API, so each keystroke is a network round-trip and needs Atlassian credentials stored in Raycast. This extension searches a local SQLite mirror maintained by the Gadak app instead:
The trade-off is the requirement: it only makes sense if you use Gadak, and results are as fresh as the mirror's last sync (about a minute behind Jira in normal use). If you want to query Jira's API directly from Raycast without the app, the existing Jira extensions are the right tool.
On a machine that installed gadak the documented way, this extension needs no preferences. It resolves the binary in this order (first existing path wins):
/opt/homebrew/bin/gadak (Homebrew, Apple silicon)/usr/local/bin/gadak (Homebrew, Intel)/Applications/Gadak.app/Contents/Resources/bin/gadak (the app bundle)Raycast's Node process does not inherit the user's shell PATH, so a gadak
that only works in Terminal is not found that way.
brew install --cask midagedev/tap/gadak (app plus CLI), or
brew install midagedev/tap/gadak-cli (CLI only)gadak init && gadak sync)Without a binary, the command offers that install command and the install guide. Search does not create a mirror.
| Preference | Default | What it does |
|---|---|---|
| Gadak Binary | empty | Absolute path. Empty uses the discovery list above. |
| Profile | empty | Passed as gadak --profile. Empty uses gadak's default profile and a deeplink with no /w/ segment. |
| Show search latency | off | When on, the results header includes the search time in milliseconds. |