Check identity and scope.
Confirm the user, workspace, and granted capabilities. The default ask/read grant does not automatically grant all tool calls.
Developers / MCP
Company context, in your AI client.
Use Heidi’s remote MCP endpoint to connect authorized clients to company knowledge and available tools.
Choose a supported authentication flow.
Compatible remote clients can use Heidi’s OAuth sign-in and consent flow. Token-based integrations can use an authorized Heidi API token with the required scopes.
Configure the connection in your client’s remote MCP settings. The exact setup depends on the client and version. Verify the identity and tool surface before using company data.
https://mcp.falkster.ai/mcpAuthorization: Bearer <your Heidi API token>Use the client’s secure credential mechanism. Never commit tokens to a repository or paste them into shared chat.
Verify the connection.
Discover available tools.
Inspect the tools exposed by the authenticated connection. Availability depends on permissions, scopes, integrations, and workspace configuration.
Test a bounded question.
Ask for company context with supporting sources. Verify the result before expanding to workflows that use actions or write data.
Build the next version of your company
Start with one outcome.
Think company-wide.
Bring a workflow that crosses departments. Let’s map the knowledge, controls, and evidence a meaningful rollout needs.
Discuss your rollout