Langdock is an enterprise AI platform with model-agnostic chat, an agent builder and workflow automation, made in Berlin and sold per user seat with add-ons.
Langdock gives a company's whole staff access to large language models through one workspace that the IT team controls. Employees chat with models from OpenAI, Anthropic, Google and other providers, build agents for repeated tasks and automate processes behind single sign-on. Langdock GmbH, based in Berlin, builds and owns the platform and has developed it since September 2023. The base subscription, Chat & Agents, pairs the chat with the Agent Builder, where a user sets an agent's instructions, model, knowledge bases and actions. A parent agent can hand tasks to subagents, and Agent Evals runs test sets against an agent before a new version goes out. Workflows adds a visual builder in which nodes, variables and integrations are dragged into place and connected. Its Guardrails node screens content for personal data, jailbreak attempts and hallucinations, while a Code node runs JavaScript or Python. Governance lets administrators review, approve, flag or disable any agent, and the API serves agents to other software, including as an MCP server. The platform suits companies rolling AI out across many departments, and the IT or security lead who has to approve that rollout. Operations teams use Workflows for processes with several steps, and developers call agents from their own applications. Each user holds a seat on Chat & Agents, in a Standard size or a Max size with more usage. Workflows is priced per workspace, Governance per user, and model calls through the API by the token, while larger rollouts sign an Enterprise contract quoted by sales. Internal knowledge search is one job, where an agent answers staff questions from documents held in Confluence, SharePoint or Google Drive. Recurring tasks become shared agents, so colleagues reuse one tested set of instructions instead of rewriting prompts. Workflows start on a schedule, from a form, from a webhook call by another system or from an event in a connected integration. When a step needs sign-off, Human in the Loop pauses the run until a person approves, rejects or fills in a form, and the workflow then continues on the matching branch.
Pricing
View plans-
Business Standard seat
$29 per month
EUR 23.20 per user a month billed annually; EUR 29.00 billed monthly; excl. VAT
- Up to 1,000 users; larger workspaces move to Enterprise
- AI models included, all features included
- SSO, SCIM and SAML
- Workflows free tier included: 2,500 runs a month, 30-day run retention, AI usage billed at API rates
- 7-day trial with EUR 5 of AI model credits, no credit card required
-
Business Max seat
$119 per month
EUR 99.00 per user a month billed annually; EUR 119.00 billed monthly; excl. VAT
- 5 times the usage of the Standard seat
- Up to 1,000 users, offered beside the Standard seat on the Business plan
-
Enterprise
Custom pricing
- Quote only, for more than 1,000 users
- Dedicated deployment possible, cloud or on-premise from 1,000+ seats
- Custom support
-
Workflows Business add-on
Custom pricing
- EUR 449 per workspace a month billed annually; EUR 539 billed monthly; excl. VAT
- 40,000 workflow runs a month; unused runs expire each billing cycle
- Needs an active Chat & Agents subscription for all users
- AI model usage billed at API token rates
- Up to 2,000 steps per workflow by default under fair usage; higher limits on request
Features
- Included: Visual agent workflow canvas
- Included: Custom tool and function definitions
- Included: Model provider choice per agent
- Included: Retrieval-augmented knowledge bases
- Included: Multi-agent handoff and delegation
- Included: Conversation state and checkpointing
- Included: Structured output schema validation
- Included: Input and output guardrails
- Included: Human-in-the-loop approval steps
- Included: Built-in execution tracing
- Included: Prompt versioning in the builder
- Included: Agent evaluation and test suites
- Included: MCP server support
- Included: Webhook and event triggers
- Included: Code-first SDK
- Included: Deploy an agent as an API endpoint
- Not included: Embeddable agent chat widget
- Included: Self-hosted agent runtime
Additional Features
- Model-agnostic chat across 40+ models
- Agent Builder with instructions, model settings and actions
- Knowledge bases attached to agents
- Subagents that take delegated tasks
- Agent Evals test sets with AI judge, tool and keyword checks
- Agent version history and workflow version rollback
- Workflows visual builder with nodes and variables
- Human in the Loop approve, reject or form steps
Best for
- Companies rolling AI out to every employee under central controls
- IT and security leads who need SSO, SCIM and ISO 27001 or SOC 2 Type II
- Teams that want models from several providers in one workspace
- Operations teams automating processes that need human sign-off
- Developers calling company agents from their own apps over an API
- Organisations above 1,000 users needing a dedicated or on-premise deployment
- Teams ready to pay per seat after a 7-day trial, not individuals
Use cases
- Giving every employee governed access to several AI models
- Building a shared agent for a recurring team task
- Answering staff questions from internal documents through a knowledge base
- Routing a workflow through a person's approval before it continues
- Starting an automation from a webhook call by another system
- Screening AI output for personal data inside a workflow
- Testing an agent against a test set before publishing a new version
- Splitting a request across specialised subagents
- Calling a company agent from an internal app over the API
- Reviewing and approving agents before staff can use them
- Rolling a workflow back to an earlier published version
Screenshots & Videos
Explore Langdock in action
User Reviews
Compare with Similar Software
Click to compare side by side with Langdock
