How does Langdock handle permissions when connected to SharePoint, Confluence, or Jira?

Native connectors use per-user OAuth 2.0 authentication and mirror the source application's permissions one-to-one. In addition, workspace sensitivity-label policies can block restricted SharePoint files from being indexed or read.

Should we build a custom integration via the Integrations API or an MCP server?

Use the Langdock Integrations API when you want lightweight, workspace-managed actions and triggers deployed from CI/CD. Build a dedicated MCP server when you want reusable tool definitions across Langdock, Cursor, Claude, and internal agent runtimes, or when tools must execute inside your private network perimeter.

Can Langdock work alongside self-hosted open-source tools like n8n, Nextcloud, and Keycloak?

Yes. We frequently wire Langdock to Keycloak for OIDC/SAML login, trigger self-hosted n8n workflows from Langdock actions, and index curated documentation from Nextcloud or OpenProject.

Do you support dedicated and self-hosted Langdock enterprise deployments?

Yes. We support both Langdock's EU cloud workspaces and dedicated enterprise deployments (using `/api/public` and `/api/public/mcp` endpoints), including network peering and custom model gateways.

Langdock consulting, enterprise integrations & MCP development

Moving Langdock from a safe chat interface to an integrated operational layer

Adopting Langdock gives teams a GDPR-compliant, model-agnostic AI workspace out of the box. The real productivity leap happens when assistants and workflows connect securely to the systems where your company knowledge and processes actually live—SharePoint, Confluence Data Center, Jira, HubSpot, ERP databases, and internal microservices. We engineer those integrations via OAuth permission mirroring, custom Model Context Protocol (MCP) servers, and Langdock's Integrations API.

What we deliver

Custom MCP servers & Langdock Agent MCP integration

Build and host remote HTTP/SSE MCP servers that expose internal databases and APIs as up to 60 governed tools per integration—or expose your Langdock workspace agents (`find_agent`, `ask_agent`) to Cursor, Claude Desktop, and internal apps via `/mcp`.

Programmatic Integrations API (Actions & Triggers)

Develop versioned custom integrations against `https://api.langdock.com/integrations/v1` (or `/api/public` on dedicated deployments) with sandboxed JavaScript actions, `requiresConfirmation` guardrails, and event triggers.

Enterprise rollout, SSO/SCIM & source-permission RAG

Configure Keycloak, Entra ID, or Okta SSO/SCIM provisioning, SharePoint sensitivity-label enforcement, and OAuth-scoped connectors for Confluence, Jira, HubSpot, and Salesforce so users never query data they lack permission to read.

Engineering standards in production

How we keep enterprise AI integrations secure and maintainable

Source-system permission inheritance

Every search and action authenticates via user-scoped OAuth 2.0 or explicit service roles—respecting SharePoint sensitivity labels, Jira project permissions, and CRM access boundaries.

Confirmation gates on write actions

Read-only lookups run smoothly in chat, while state-changing actions (creating Jira issues, updating HubSpot deals, triggering ERP workflows) enforce `requiresConfirmation` and schema validation.

Pairing Langdock with n8n & backend services

Keep conversational UX and knowledge assistants in Langdock while delegating heavy multi-system orchestration, retries, and webhook processing to self-hosted n8n or backend queues.

When engagement makes sense

Connecting proprietary or on-premise systems

When standard SaaS connectors are not enough and you need custom MCP servers or Integrations API bridges into internal ERPs, PostgreSQL databases, or self-hosted Jira/Confluence Data Center.

Regulated enterprise rollout & governance

When your security team requires clean SSO/SCIM group mapping, model-routing policies by data classification, and auditable API key scopes (`INTEGRATION_API`, `ASSISTANT_API`).

High-leverage departmental agents

When sales, engineering, legal, or support teams need domain-specific assistants that combine curated knowledge folders, live CRM/ticket lookups, and deterministic workflow triggers.

Related guides & services

FAQ

How does Langdock handle permissions when connected to SharePoint, Confluence, or Jira?

Native connectors use per-user OAuth 2.0 authentication and mirror the source application's permissions one-to-one. In addition, workspace sensitivity-label policies can block restricted SharePoint files from being indexed or read.
Should we build a custom integration via the Integrations API or an MCP server?

Use the Langdock Integrations API when you want lightweight, workspace-managed actions and triggers deployed from CI/CD. Build a dedicated MCP server when you want reusable tool definitions across Langdock, Cursor, Claude, and internal agent runtimes, or when tools must execute inside your private network perimeter.
Can Langdock work alongside self-hosted open-source tools like n8n, Nextcloud, and Keycloak?

Yes. We frequently wire Langdock to Keycloak for OIDC/SAML login, trigger self-hosted n8n workflows from Langdock actions, and index curated documentation from Nextcloud or OpenProject.
Do you support dedicated and self-hosted Langdock enterprise deployments?

Yes. We support both Langdock's EU cloud workspaces and dedicated enterprise deployments (using `/api/public` and `/api/public/mcp` endpoints), including network peering and custom model gateways.

Discuss your Langdock rollout and integration architecture

We scope SSO, custom MCP servers, API integrations, and high-value departmental agents with your team.

Langdock enterprise rollout, custom integrations, and MCP servers

Connect your company AI workspace to SharePoint, Confluence, Jira, HubSpot, and internal APIs—with strict permission inheritance, SSO, and custom MCP tooling.