No user-submitted related servers found.
Scored across 143 tools
Nearly every tool is namespaced by domain (bots_, flows_, dialogs_, broadcasts_, knowledge_, requests_, counterparties_, extensions_, market_, etc.) and descriptions actively draw boundaries, e.g. bots_journal vs audit_list and dialogs_history vs dialogs_export. Some residual overlap remains since several *_save tools upsert (create+update) and there are clusters of similar read tools (bots_webhook_info/check/rotate, notifications_*), but an agent can reliably pick the right tool.
Names follow a predictable snake_case resource_action pattern (bots_list, flows_get, dialogs_reply, market_install) with sub-resources consistently grouped (bots_commands_apply, requests_types_save). A few outliers break the pattern (whoami, operbots_catalog, operbots_login/logout), but they are the exception and remain readable.
At 143 tools this is far beyond the practical scope of a single MCP server; even though the underlying panel domain is genuinely broad (bots, flows, dialogs, broadcasts, knowledge, AI, tickets, counterparties, extensions, market, audit), the count will strain context and tool selection, and several domains would be better split or trimmed.
The surface covers full lifecycles across every apparent sub-domain: CRUD for cases, bots, flows (with versions/import/export), dialogs, broadcasts, knowledge bases, AI providers, requests/tickets and their types, counterparties, extensions, plus members, roles, invites, sessions, account, audit, notifications and a market. Few obvious gaps remain, and edge operations like reindex, retry, reconcile and reset_stage are included.