Tough Tongue AI

Get detailed collection information with org context support. Access is granted if: - User is the creator of the collection - User has permissions to the collection (via Permission system) - Org has permissions to the collection (if in org context)

get_collection

Deployment settings safe for any caller. No login. A PAT does not change the body.

get_public_config

Get detailed session information including transcript and evaluation data. The only public route returning a session transcript — /v2/sessions/batch omits it.

get_session

Fetch specific sessions by ID. Much faster than listing with filters.

get_sessions_batch

Get personal balance or organization shared and member-quota balances. Active organization members may read the shared organization balance. In a quota-enabled organization, an allocated caller quota is included; its absence does not hide the shared balance.

get_balance

Internal to Tough Tongue Studio's Edit-with-AI chat: shows the user 1-4 structured clarifying questions as a card and ends the turn; the answer arrives as the next user message. Has no effect in other clients — ask in plain chat instead and do not call this tool.

int_ask_user_question

Get legacy current or selected-version Scenario content under consumption access. Prefer `v3_get_scenario_version` for a known Scenario ID's current content or exact archived version. Retain this route for legacy response compatibility.

get_scenario

List full legacy Scenario records with organization-context support. Use `v3_list_scenarios` by default for current inventory, predicate filtering, or counts. On MCP, use this legacy read only for case-insensitive free-text search or its established full summary response shape. Direct REST callers can additionally use dynamic `meta_*` filters. Returns: - Scenarios created by the user (with org filtering) - Scenarios where the user has permissions (with org filtering) - Scenarios where the org has permissions (if in org context) - When is_featured=true: globally featured public scenarios (org ignored) Filters: - search: case-insensitive substring match on name + description - meta_<key>=value (exact match on user_metadata) - meta_$gt_<key>=value, meta_$gte_<key>=value, meta_$lt_<key>=value, meta_$lte_<key>=value

list_scenarios

List enriched legacy Sessions with evaluation and improvement results. Use `v3_list_sessions` by default for current inventory, exact Scenario/status filtering, or counts. Use this legacy read only for person, date, metadata, learning, or enriched-result filters that V3 does not publish. Response includes: - evaluation_results: Full evaluation (score, report_card, strengths, weaknesses) - improvement_results: Improvement areas, action items, resources Filters: - scenario_id, user_email, is_org, page, limit Date filters (convenience): - from_date / to_date: shorthand for updated_at range. Date-only spans 00:00:00 through 23:59:59.999999 UTC. Date filters (advanced — $op_<field> convention, same as metadata): - $gte_updated_at, $gt_updated_at, $lte_updated_at, $lt_updated_at - $gte_created_at, $gt_created_at, $lte_created_at, $lt_created_at - Timestamps accept Z or a numeric offset; no timezone means UTC. Both styles can be used together; advanced filters take precedence on overlap. Metadata filters: - meta_<key>=value (exact match on user_metadata) - meta_$gt_<key>=value, meta_$gte_<key>=value, meta_$lt_<key>=value, meta_$lte_<key>=value

list_sessions

List all collections through the public API with org context support. Returns: - Collections created by the user (with org filtering) - Collections where the user has permissions (with org filtering) - Collections where the org has permissions (if in org context)

list_collections

List meeting bot records with pagination. Filters: - status: pending, scheduled, in_call_recording, call_ended, failed - scenario_id: Filter by specific scenario - from_date/to_date: ISO date/time; date-only spans the relevant UTC day boundary - page/limit: Pagination (default: page=1, limit=50, max=500)

list_meeting_bots

List organizations the authenticated PAT user belongs to. Each item's `id` is the value the client should send back as the `X-TT-ORG` header to scope subsequent API calls into that organization. Omitting the header (or sending an empty value) keeps the call in personal context.

list_organizations

List SIP call records with pagination. Filters: - status: pending, scheduled, in_call_recording, call_ended, failed - scenario_id: Filter by specific scenario - batch_id: Filter by batch ID - from_date/to_date: ISO date/time; date-only spans the relevant UTC day boundary - page/limit: Pagination (default: page=1, limit=50, max=500)

list_sip_calls

List configured SIP trunks for the caller (org-scoped when X-TT-ORG is set). Use this to discover the required `sip_trunk_id` before dispatching outbound calls with `POST /sip/call` or `POST /sip/batch`. Outbound calls run on `outgoing` trunks. Credentials (auth username/password, LiveKit IDs) are never returned.

list_sip_trunks

Fill in missing post-session results (analysis, extraction, learning) in the background. Only jobs the scenario configures and whose results are still missing run; existing results are never overwritten. Use it to retry after a failed or skipped analysis. New results can fire the scenario's configured post-session webhooks. Response is always 200/OK — processing is async. Poll `GET /sessions/{id}` and check `post_session_status.state` (`running` → `idle` on success, `failed` on error) or the presence of `evaluation_results` / `extraction_results` to detect completion. Nothing configured, nothing missing, or a run already in progress → silent no-op.

post_process_session

Schedule meeting bot(s) via API token. Supports multiple URLs (newline-separated). The bot joins the meeting as a visible named participant and posts a chat message to everyone on joining, sent under its display name. Choose a bot_name that identifies it as an AI participant so attendees know they are talking to an agent — e.g. "Ava (AI Coach)", "Tough Tongue AI Notetaker". Requires: - Paid plan - User must have EDIT+ permission on the Scenario - Org context: If X-TT-ORG header present, scenario must belong to that org

schedule_meeting_bot

Call once before the first Tough Tongue MCP action in a conversation unless the guide is already in context. Returns the V3-first read policy and MCP quickstart, also available as the ttai://guide/mcp-agent resource.

callme_before_using_tough_tongue_mcp

Update an existing scenario. `id` is required; only provided fields change. Pass org_id when the scenario belongs to an organization. Use create_scenario for new scenarios.

update_scenario

Get agent-safe effective plan capabilities for the current PAT scope.

v3_get_entitlements

Get one safe resource summary from the current PAT scope.

v3_get_resource

Get current Scenario content or one version by exact ID or name. This is the default MCP detail read when a Scenario ID is known. Omit both selectors for current HEAD with normal VIEW+ access. Selecting a version inherits the same consumption access; current HEAD wins when a name was reused.

v3_get_scenario_version

List generic V3 resource types, fields, operations, and semantics.

v3_list_resource_types

List one safe, searchable cursor page in the current PAT scope.

v3_list_resources

List current HEAD and every archive's IDs, labels, and safe details for EDIT+.

v3_list_scenario_versions

List current Scenario heads in the active personal or org scope. This is the default MCP read for Scenario inventory, predicate-filtered pages, and exact counts. Use legacy `list_scenarios` only for free-text search.

v3_list_scenarios

List a compact cursor page of raw Sessions in the current PAT scope. This is the default MCP read for Session inventory, exact Scenario/status filters, and counts. Use legacy `list_sessions` only for person, date, metadata, learning, or enriched-result filters.

v3_list_sessions