The API Report CardAPI Index
Intercom

Intercom API

Customer Messaging / AI Helpdesk / Conversational Support · intercom.com

Intercom offers a self-serve REST API at developers.intercom.com covering contacts, conversations, tickets, articles, and events, with webhooks and Canvas Kit apps. Access tokens are workspace-scoped; OAuth serves public apps. Throttling near 83 requests per 10s is the real ceiling.

Last verified: August 2026Marketing & SalesMCP READY
API GRADE
A
VERIFIED AUG 2026

SCORECARD

ExistenceGOODDocumented REST API at api.intercom.io spanning contacts, conversations, tickets, articles, and events.
AccessGOODSelf-serve access tokens from the Developer Hub; higher rate limits require contacting support and often a plan upgrade.
CoverageGOODContacts, companies, conversations, tickets, and help center content; full message bodies need the paid S3 export.
AuthGOODWorkspace-scoped access tokens for private apps, OAuth 2.0 with granular scopes for App Store listings.
Docs & DXGOODOfficial Node, Ruby, PHP, Java, and .NET SDKs plus Canvas Kit for in-inbox apps and webhook topic mapping.
StabilityGOODVersioned endpoints back to 1.4 remain supported, though running many live versions creates inconsistent field shapes.
MCPGOODIntercom runs an official hosted MCP server at mcp.intercom.com with 13 tools covering conversations, contacts, companies, and Help Center articles, mostly read-only with article write access. OAuth setup instructions are documented on developers.intercom.com for US and EU hosted workspaces. developers.intercom.com VERIFIED 2026-08
Supergood turns hard-to-integrate enterprise software into clean REST APIs and MCP tools: stable endpoints, normalized JSON, managed auth.

Frequently asked questions

Intercom scores A on Supergood's API Report Card. Intercom offers a self-serve REST API at developers.intercom.com covering contacts, conversations, tickets, articles, and events, with webhooks and Canvas Kit apps. Access tokens are workspace-scoped; OAuth serves public apps. Throttling near 83 requests per 10s is the real ceiling.

Tried to integrate with Intercom?
SOURCES
Practical rate limit of ~83 requests per 10 seconds per token chokes bulk operations; documented headline of 10k/min per app is rarely the binding constraint github.com
~500 user updates per minute cap makes bulk user sync from external CDPs/data warehouses painfully slow github.com
429 throttling on conversation list endpoints during routine bulk fetches; backoff strategies break under sliding 10-second windows community.intercom.com
Higher rate limits are not self-serve, require contacting Intercom support and often a plan upgrade developers.intercom.com
Conversation message bodies and attachments are not available via standard CSV export, full historical content requires the paid Amazon S3 / Google Cloud Storage data export feature intercom.com
Multiple API versions (1.3, 1.4, 2.0-2.13) supported simultaneously create migration drag and inconsistent field shapes developers.intercom.com
OAuth scope granularity forces apps to over-request permissions for common multi-resource workflows developers.intercom.com
Fin Agent API and the standard Intercom REST API are separate surfaces with separate auth flows, integrators wanting both pay double the integration cost developers.intercom.com
Fin per-resolution billing produces unpredictable monthly cost; one widely-shared Reddit post reported a $4k/mo bill jumping to $9k after enabling Fin kommunicate.io
'Assumed resolution' billing counts conversations as Fin-resolved even when a human takes over; Intercom promised a fix at Built For You then walked it back myaskai.com
12-month minimum contracts with no early-termination, customers report being forced to pay through the remainder of the term g2.com
Contact-volume billing surprises customers because Intercom counts contacts shown a form/Messenger, not just those who interact, pricing scales with traffic not engagement trustradius.com
Price escalation has no published schedule; Fin can exceed prior human-agent cost-per-resolution at scale clearfeed.ai
Headline $29/seat pricing is wildly disconnected from real-world cost once Fin outcomes, premium channels, and add-ons stack up costbench.com
Reporting depth is weaker than Zendesk Explore for high-volume queue analytics; many teams pipe data to Looker/Hex via the S3 export gartner.com
CSV exports drop all conversation message bodies and attachments, full content export requires the paid S3/GCS data export feature clonepartner.com