Plain + clariBI
Connect Plain through clariBI's MCP catalog. Read-only OAuth, no CSV exports, automatic dashboards on support volume, first-response time, theme breakdowns, and customer health.
Native Plain connector via the MCP catalog.
Connect with OAuth 2.0 + PKCE, the AI engine inherits your Plain user's permissions, ask questions across support + revenue + product data in plain English. The setup is the same one-click flow as Stripe, Calendly, or HubSpot.
What is Plain?
Plain is a modern B2B customer-support platform built around a thread-based model, with API-first workflows that integrate cleanly with the rest of a product team's stack.
clariBI reaches your Plain workspace through the official Plain MCP server (generally available since March 2026): threads, customers, tenants, help-center articles, labels, and workspace data. Reads inherit the connected user's Plain permissions. Write tools (reply, assign, snooze, mark-done) are filtered out at the catalog layer — clariBI is analytical, not actuating.
Learn more at plain.com
Connect Plain in three steps
Open Data Sources in clariBI
In clariBI, go to Data Sources → Add Data Source → MCP Servers tab. Find Plain in the catalog and click Connect.
Authorize with Plain (OAuth + PKCE)
A Plain OAuth popup opens with your Plain user's permissions inherited. No API keys to manage. Approve. clariBI stores credentials Fernet-encrypted at rest.
Ask questions, get dashboards
clariBI surfaces ticket volume, first-response time, customer-tier breakdowns, and theme mix. Ask "Which customer tiers drove the most tickets in November?" or "What's the median first-response time on bug-tagged threads?" — the AI engine queries Plain alongside Stripe, HubSpot, and your other sources.
What you can analyze
Out of the box with the Plain MCP connector:
- Thread volume over time (by customer tier, theme, label)
- First-response time and SLA compliance
- Customer concentration: which accounts drive the most volume
- Theme breakdown (bug, feature request, billing, onboarding)
- Unassigned-thread aging and queue health
- Cross-source: support volume vs Stripe churn risk
Plain is one of 60+ vendors in the clariBI MCP catalog. See the full list →
What the AI engine can call
Plain exposes 30 granular tools across five domains. clariBI's planner gets the read surface only — write actions (reply, assign, snooze, mark done, upsert) are filtered out at the catalog layer before the planner ever sees them.
Threads (the support conversations)
Fetch and search support threads. Pull full details for a single thread including its replies, labels, assignee, status, and customer reference.
Reads enabled: fetch_thread, search_threads, get_thread_details
Customers (who's writing in)
Fetch customer profiles by email or ID, search by attributes, and list all threads associated with a customer for retention and concentration analysis.
Reads enabled: fetch_customer, search_customers, list_customer_threads
Tenants (the customer's company)
Fetch tenant records with the firm-level metadata Plain tracks (plan tier, ARR, contract value). The join key for tying support volume to revenue tier.
Reads enabled: fetch_tenant, search_tenants
Help Center & Workspace
Read help-center articles by ID or slug; surface workspace metadata and label definitions. Lets the AI engine ground tickets against the published self-serve content.
Reads enabled: get_help_center_article, list_labels, get_workspace
Cross-source questions Plain unlocks
Support data on its own answers volume + response-time questions. Joined to revenue and CRM, it answers questions about retention risk, support load per ARR dollar, and theme-vs-churn correlation.
Plain × Stripe
"Which Stripe customers churned in Q3 had open bug-tagged threads in the 30 days before cancel?"
Planner: list_canceled_subscriptions on Stripe → lookup_mcp_entities on Plain customer emails for each → filter threads by label=bug + time window → rank by lost revenue.
Plain × HubSpot
"What's our support-cost-to-ARR ratio for each customer tier?"
Planner: aggregate thread count + median first-response time on Plain by customer → join HubSpot tier & ARR → compute ratio, top + bottom outliers per tier.
Plain theme tracking
"Which support themes have spiked the most week-over-week and which features do they map to?"
Planner: search_threads by label → week-over-week deltas → cross-reference with help-centre articles & product analytics for the affected feature areas.
Plain connector FAQ
Which Plain plan do I need?
Any Plan tier that exposes the public API. Plain's MCP server went generally available in March 2026 — if you can reach the API today, the MCP server is enabled on your workspace.
Does clariBI inherit my Plain permissions?
Yes. Plain MCP uses OAuth 2.0 with PKCE and ties to your specific Plain user. Whatever threads, customers, tenants, and labels your account can see in Plain, clariBI can read. Whatever's hidden by Plain's role model stays hidden in clariBI.
Can clariBI reply to threads, assign tickets, or change status?
No. Plain exposes write tools (reply, assign, snooze, mark done, upsert) but clariBI's catalog layer filters them out before the planner ever sees them. The connector is strictly analytical.
Are API keys involved anywhere?
No. Plain's OAuth + PKCE flow handles auth end-to-end. No tokens, secrets, or API keys are stored in clariBI by the user — DCR-style credentials are issued and managed automatically.
My Plain workspace is US-hosted. Does that matter?
No. The remote MCP endpoint at mcp.plain.com/mcp routes to the right region transparently. Both US-hosted and EU-hosted workspaces work without additional configuration.
Start your free trial and connect Plain in a minute
OAuth in, read-only, dashboards in seconds.
Start free trial