Tauri command reference
The UI talks to the Rust core through Tauri commands (invoke). This is the
app’s real API surface. All commands return an error string on failure.
Conversation
Section titled “Conversation”get_status
Section titled “get_status”Provider and readiness snapshot. Returns:
{ providers: { id: string; configured: boolean; reachable: boolean | null; model: string }[]; ready: boolean; onboarding: string | null; // setup guidance when no provider is available message_count: number; fact_count: number;}chat_send
Section titled “chat_send”Send a user turn; get the assistant reply.
- Args:
{ text: string } - Returns:
{ content, provider, model, cached: boolean, confidence: number | null } - Errors: empty message; all providers failed; no provider configured.
get_history
Section titled “get_history”- Args:
{ limit?: number }(default 200) - Returns:
{ id, role, content, created_at }[]in chronological order.
Telemetry
Section titled “Telemetry”get_telemetry
Section titled “get_telemetry”Live machine + app vitals. Returns:
{ cpu_percent: number; mem_used: number; mem_total: number; uptime_secs: number; note_count: number; message_count: number; fact_count: number }The first call reports 0 CPU (a prior sample is needed); it settles by the next poll.
Skills
Section titled “Skills”| Command | Args | Returns |
|---|---|---|
list_skills | — | SkillManifest[] |
save_skill | { name, description, code, test } | SkillManifest |
author_skill | { request } | { manifest, attempts, passed } |
test_skill | { name } | SkillManifest |
run_skill | { name, input } | string (output) — errors if flagged |
SkillManifest: { name, version, description, created_at, updated_at, test_status }
where test_status is { status: "passed" } or { status: "failed", detail }.
See the skill manifest reference.
| Command | Args | Returns |
|---|---|---|
list_notes | — | string[] (slugs) |
read_note | { name } | string |
save_note | { title, content } | string (slug) |
Memory, reflection, replay
Section titled “Memory, reflection, replay”| Command | Args | Returns |
|---|---|---|
get_events | { limit?: number } | Event[] |
list_insights | { limit?: number } | Insight[] |
reflect_now | — | Insight[] (new lessons) |
reflect_if_due | — | number | null (count, or null if not due) |
undo_event | { eventId } | string (human-readable outcome) |
replay_audit | — | ReplayReport |
export_memory | — | JSON (messages, facts, insights, events, notes) |
wipe_memory | — | — (clears memory + event log; keeps notes) |
ReplayReport: { matched, missing_in_db, extra_in_db, deterministic }. See the
event log reference.