Stack Ai

@stackai what actions are available for the Slack provider and what parameters does the 'send message' action require?

Based on the Stack AI provider registry, the Slack provider offers 8 actions including 'send_message', 'create_channel', 'list_channels', and 'get_user_info'. The 'send_message' action requires three parameters: 'channel' (string, the channel ID or name), 'text' (string, the message content), and an optional 'thread_ts' (string, for replying in threads). The action returns a message object with timestamp and confirmation status. This provider also supports 2 triggers for incoming messages and reactions.

This example demonstrates Stack AI's provider introspection capabilities — useful when building automation workflows and you need to discover what integrations are possible. Requires a valid API key with read access to the provider catalog. The actual available actions depend on your Stack AI workspace configuration and installed providers.

@stackai check if the API is healthy and show me the current license status including expiration date

Stack AI API health check: All systems operational (status: healthy, response time: 142ms). Your Stack AI license is currently active with the Professional tier. License details: valid through March 15, 2025 (47 days remaining), supports up to 10 team members and 50 active workflows. No service degradation detected. Last health check performed at 2025-01-27 14:32 UTC.

This combines two monitoring tools to give a quick operational snapshot — helpful for troubleshooting integration issues or verifying service availability before deploying new workflows. The health check runs in real-time; license data reflects your actual subscription state and may trigger renewal reminders as expiration approaches.

@stackai I need to configure a new Google Sheets connector — what configuration parameters are required and what's the expected schema format?

The Google Sheets connector in Stack AI requires the following configuration parameters: 'credentials' (OAuth2 token object with access_token and refresh_token), 'spreadsheet_id' (string, the Google Sheets document ID from the URL), and optional 'sheet_name' (string, defaults to first sheet). The connector schema expects JSON format with nested credential objects. Output schema includes columns as key-value pairs where keys are column headers and values are cell contents. The connector supports both read operations (fetch rows, get cell values) and write operations (append rows, update cells).

This example shows Stack AI's schema introspection for connector setup — critical when integrating new data sources into workflows. The actual configuration process happens in the Stack AI dashboard; this MCP provides the schema documentation to guide setup. OAuth credentials must be obtained through Google's authorization flow separately.

When Stack AI wins for mapping third-party tool chains

A 6-person ops team runs 15 SaaS tools and needs to automate handoffs between them—Slack to Linear, Stripe to Notion, Zendesk to Airtable. Stack AI's MCP is the right call here because its 22 tools are all about introspecting provider schemas and action inputs. Your team can query "what parameters does the Stripe webhook action need" or "show me the output schema for Zendesk ticket creation" without leaving the AI workspace. This works best when you're in the design phase of automation, not runtime execution. If you need to actually trigger those workflows in real-time, you'll hit the MCP's read-only boundary and need Stack AI's full platform. For teams spending 3+ hours a week debugging integration configs, this MCP cuts that discovery loop from browser tabs to a single conversational thread.

When this MCP replaces manual status checks

A 3-person support engineering team fields escalations about integration failures across a dozen third-party APIs. Stack AI's MCP gives them a single "Check Health" tool to verify Stack AI's own API status, plus schema lookups to confirm whether a connector's expected inputs have changed upstream. This is a narrow win: it only monitors Stack AI's service layer, not the underlying providers themselves. Use this when your team already runs Stack AI workflows in production and needs fast triage during incidents. If your integrations live outside Stack AI's ecosystem, the health check won't help. The real value is pairing health status with schema introspection—when a webhook starts failing, you can immediately ask "did the Stripe action input schema change" and get a JSON diff in seconds. For teams handling 10+ integration tickets per week, that's 2 hours saved.

When this MCP speeds up new-hire connector setup

A solo integration engineer at a 12-person startup is onboarding two new hires who need to learn how Stack AI's 40+ provider connectors work. The MCP's "Get Provider Details" and "Get Connector Type Schema" tools let the new hires ask "what does the Salesforce connector require" or "show me all actions for the HubSpot provider" without reading static docs. This works when your team is already committed to Stack AI as the integration backbone. If you're evaluating multiple iPaaS platforms, the MCP won't help you compare—it only introspects Stack AI's own provider catalog. The buying threshold is simple: if onboarding takes more than 4 hours per engineer because they're context-switching between Notion docs and the Stack AI dashboard, this MCP collapses that into conversational lookup. Best for teams scaling from 1 to 3+ integration engineers in a quarter.