TehProf Support

What TehProf Support does

TehProf Support MCP server exposes a TehProf Support workspace to MCP-compatible AI clients. The platform is a multi-tenant SaaS helpdesk for IT service companies, and the MCP interface covers both read and write operations subject to the tenant’s plan and permissions.

The available functionality includes full-text knowledge-base search, article retrieval, onboarding and pricing information, tenant and system inspection, settings access, operator lists, ticket queries, ticket creation, replies, status and priority changes, channel status, analytics, automation rules, and outbound webhook inspection. Ticket creation can synchronize a customer company or contact using stable external IDs and can avoid duplicates with an external reference. The supplied tool catalog also includes sandbox ticket demonstrations and Bitrix24-related CRM operations in the Business tier.

This fits support teams that want an assistant to find internal answers, review ticket context, perform permitted ticket work, or inspect operational metrics without building a separate integration layer.

How it works

The hosted service uses Streamable HTTP with JSON-RPC 2.0 and SSE-related MCP transport behavior. Clients connect to https://support.tehprof.kz/mcp; the server then communicates with the TehProf Support backend. A discovery document is available at /.well-known/mcp.json, where compatible clients can find capability and authentication information.

Anonymous requests expose the free tool set, including knowledge, onboarding, demo, pricing, system, and settings functions. Starter, Pro, and Business plans add progressively more capabilities. The server hides tools that are not available to the current plan, so an agent does not receive unavailable operations through tools/list.

Authenticated requests use a tenant API key in an Authorization: Bearer header. Write actions require the appropriate write scope, and calls are resolved against the tenant associated with the key. The documented security model includes tenant checks, rate limiting, audit logging, TLS, and server-side secret storage.

Setup and configuration

TehProf Support MCP server can be used without local installation through the hosted endpoint. Add a Streamable HTTP MCP entry to the client configuration and include the bearer token when authenticated tools are needed. Create the key from the TehProf Support settings area under API Keys.

Self-hosting requires Node.js 20 or newer. Clone the repository, install dependencies, build the project, and start dist/index.js. Set SUPPORT_BACKEND_URL to the Support backend address. The example startup also sets MCP_PORT and MCP_HOST; the documented example uses port 8101 and host 127.0.0.1.

The README names Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, ChatGPT Custom GPTs, and custom MCP SDK consumers as supported connection options. ChatGPT configuration is provided separately in the repository.

Tools and capabilities

• Search, read, and categorize knowledge-base articles.
• List, inspect, create, reply to, and update support tickets according to plan and scope.
• Review dashboard KPIs, SLA compliance, and per-operator performance.
• Inspect channels, tenant settings, operators, system status, and tenant metadata.
• Create and toggle automation rules, then inspect outbound webhooks.
• Use sandbox ticket workflows, pricing tools, onboarding guidance, and Bitrix24 CRM tools where the applicable tier enables them.

Limitations and notes

The TehProf Support MCP server does not provide every tool to every caller. Subscription tier determines which operations are exposed, while authentication and scopes further constrain tenant data and write actions. Anonymous access is limited to the documented free tier.

The supplied tool list describes 36 AI tools in the README, while the tier tables also document Business-tier Bitrix24 and channel-configuration tools. Availability should therefore be checked against the connected tenant’s plan and the server’s discovery response rather than assumed from a client configuration.

The hosted interface is Streamable HTTP rather than a local stdio server. Teams requiring deployment inside their own network must follow the repository’s self-hosting instructions and provide a reachable Support backend.