fairw.ai

Healthcheck. Returns the authenticated user and tenant — use to verify the connector is wired correctly. No side effects.

ping

Fetch the full content of one specific meeting from a record_id (typically obtained from `list_meetings`, `search_meetings` or `keyword_search`). Returns the AI-generated report (decisions, key points, discussion), the participants, and the action items with assignees, due dates and status. For exact verbatim quotes from the conversation use `get_transcript` instead.

get_meeting

Fetch the raw verbatim transcript of a meeting, paginated (~3000 tokens per page) with [timestamp] Speaker: text formatting. Best when the user needs an exact quote, the precise wording someone used, or the full back-and-forth. Optionally filter by speaker. Prefer `get_meeting` for a summary or decisions; only call this tool when verbatim content is required.

get_transcript

List the authenticated user's past meetings (completed recordings only). Supports filtering by date range, participants (all required, logical AND), customer domain, and title substring. Best when the user wants to list or count meetings by metadata — e.g. 'my meetings last week', 'all my calls with Acme this month', 'meetings with sarah@example.com'. Returns each meeting with record_id, title, date, duration, participants and a short summary; drill into a specific meeting via `get_meeting`. For UPCOMING calendar meetings use `list_upcoming_meetings`. For meaning-based content search use `search_meetings`.

list_meetings

List the authenticated user's upcoming calendar meetings (synced from Google Calendar / Outlook). Best when the user asks about their future calendar — 'what do I have tomorrow', 'my next meeting', 'prepare me for my call with Acme', 'what's on my calendar this week'. Returns title, start/end time, timezone, participants, agenda, video link, and whether each meeting will be recorded. Times are expressed in each meeting's own time zone with its explicit UTC offset, e.g. `2026-10-01 09:00–10:00 (Europe/Paris, UTC+02:00)`: convert them if the user is in another time zone. Combine with `list_meetings` or `search_meetings` to pull historical context about the same participants.

list_upcoming_meetings

Lexical (keyword) search across the user's meetings — exact match with typo tolerance, scoped to titles, summaries and transcripts. Best for terms that must match literally: contract references like 'SO-2024-118', order numbers, ticket IDs, precise monetary amounts, rare proper nouns, project codenames. Returns matching records with title, date, duration and a snippet where the term appears. For conceptual or meaning-based questions ('what did we say about X') prefer `search_meetings` instead.

keyword_search

Semantic search across the user's meeting transcripts. Returns the most relevant passages with their meeting context (title, date, speakers) and a record_id to drill into via `get_meeting` or `get_transcript`. Best for meaning-based questions about what was discussed, decided or raised — e.g. 'what did we say about pricing', 'who raised concerns on the roadmap', 'what did Acme push back on'. For literal exact-term lookup (contract IDs, amounts, rare proper nouns) prefer `keyword_search`. For metadata-only filtering (date range, participants, customer domain) prefer `list_meetings`.

search_meetings