Langdock
Free
Langdock is an enterprise-level
Langdock tool analysis
Core parameters and statistics
| Parameters | Official public information |
|---|---|
| Product positioning | Enterprise AI Adoption platform, covering Chat, Agents, Workflows, Integrations, API |
| Company entity | Langdock GmbH |
| Place and time of establishment | Founded in 2023 in Berlin, Germany, see Press page |
| Platform entrance | Web, iOS, Android, API |
| Model strategy | Model-independent, the official website says it can access OpenAI, Anthropic, Mistral, Google and other models |
| Integration scale | Integrations page exposes 57 integrations and 754 native actions |
| API capabilities | Completion API, Embedding API, Agent API, Knowledge Folder API |
| Default API limits | API page exposes 500 requests/min and 60,000 tokens/min |
| Security Compliance | Security Page Disclosure ISO 27001, SOC 2 Type II, GDPR, TLS 1.2+, AES-256 |
| Deployment forms | Multi-tenant SaaS, single-tenant SaaS, own cloud, local Kubernetes/Helm |
The core parameter of Langdock is not the context length of a single model, but the governance layer, knowledge layer, integration layer and model layer that enterprises need when bringing AI into the organization. It puts daily employee chat, department-level agents, process automation, internal tool connections, and unified model APIs in the same workspace. It is suitable for enterprises that require unified permissions, auditing, knowledge access, and cost control, rather than individual users who just want to experience a general chatbot.
User and market recognition
Langdock claims that it is "Trusted by over 7,000 companies" on the homepage of its official website, and also discloses 4,000+ customers, 100K+ monthly active users, 10M+ messages per month, 30+ employees and 2025 revenue growth 10x on the Press page. The two "company/customer" figures come from different pages and calibers. This article does not combine them and only cites them according to the context of the official page.
The corporate customer signal is also strong. The homepage of the official website displays customer logos such as Merck, GetYourGuide, Eppendorf, Der Spiegel, SumUp, The Economist, Axpo, UNICEF, Statista, etc.; the homepage case quotes Merck Chief Data & AI Officer's statement that Merck's "myGPT Suite (Langdock)" has become a company-level AI companion and is used by 33,000+ monthly active users. This type of information suggests that Langdock's market focus is on organizational AI promotion, enterprise knowledge, and security governance, rather than personal productivity tools.
From a positioning perspective, Langdock captures the typical constraints of European enterprises in AI adoption: data sovereignty GDPR, model selection, permission mirroring, employee ease of use and internal tool integration. Its differentiation is not just "the ability to chat", but rather enabling IT, data, legal and business departments to manage AI usage boundaries on the same platform.
Cost advantage
| Plan/Module | Official public price | Applicable objects | Description |
|---|---|---|---|
| Trial | Free | Trial Team | Pricing Page Public 7-day free trial, no credit card required, and includes 5€ AI model credits |
| Chat & Agents Business Standard | EUR 25 / user / month | SMEs and departmental teams | Save 20% with annual payment, price excludes VAT |
| Chat & Agents Business Max | EUR 99 / user / month | Higher usage team | 5x more usage marked as Standard seat on the official website |
| Chat & Agents Enterprise | Custom | 1,000+ users | Supports large-scale deployment, dedicated deployment and customization support |
| Workflows Starter | Included in Chat & Agents | Initial Automation | 2,500 workflow runs / month |
| Workflows Business | Starting from EUR 539 / workspace / month | Workflow scaling team | 40k workflow runs / month; 100k runs shown on the official website is EUR 1199 / month |
| API Add-on | Billing by model token | Developers and platform team | The API page discloses the model input/output token price, which is subject to the real-time page |
Langdock's cost advantage comes mainly from two points. First, Chat & Agents officially states “no usage-based cost component” when including AI models, which makes it easier for companies to budget for employee-level promotions. Second, API, Workflows and Chat/Agents are priced separately. Enterprises can first use fixed seats to promote daily use, and then manage high-consuming automation and model calls separately.
Note that Enterprise, single tenant, own cloud, on-premises, more workflow runs, and higher API limits are subject to confirmation with sales. When purchasing, you should focus on checking the seat size, model inclusion range, workflow running retention period, API price VAT, annual payment discount and enterprise deployment conditions.
Main functions
- Chat: A company-wide model-independent chat portal. The official website states that it allows employees to use leading models in a unified workspace and reduces Shadow IT.
- Agents: In the Agents page, Langdock defines Agents as dedicated chatbots that can be configured with specific use cases or documents. They support saving instructions, documents, and context, and are suitable for repetitive tasks and team sharing.
- Knowledge management: Agents can upload files and use knowledge folders to center answers around company facts rather than generalized knowledge.
- Workflows: used to build AI automated processes, suitable for orchestrating approval, summary, data processing, notifications and cross-system actions.
- Integrations: Officially disclosed 57 integrations and 754 native actions, and listed Google Drive, Notion, Confluence, Slack, Salesforce, Jira, Asana and other work tools.
- MCP support: Agents FAQ clearly states that Model Context Protocol is supported. Tools exposed by MCP server can be automatically discovered and converted into Langdock actions, and authentication methods such as API key and OAuth 2.0 are supported.
- API: Provides Completion, Embedding, Agent and Knowledge Folder APIs, suitable for integrating unified model access to RAG and Agent capabilities into your own applications.
- Governance & Security: Supports workspace permissions SSO/SCIM/SAML, permissions mirroring, optional analytics, subprocessor transparency and enterprise deployment options.
After combining these functions, Langdock is more like an "enterprise AI operation layer": ordinary employees use Chat and shared Agents, business teams use Workflows to automate processes, developers use APIs and integrations to connect to existing systems, and IT and security teams control risks through deployment, permissions, and compliance.
Model and version evolution
| Stage | Time | Disclosure basis | Description |
|---|---|---|---|
| Company Founded | 2023 | Press Page | Official Disclosure Langdock Founded 2023 in Berlin |
| Organization-level platformization | 2026 Current official website | Home | The official website summarizes the products into five core modules: Chat, Workflows, Agents, Integrations, and API |
| Integrations and MCP extensions | 2026 Current official website | Integrations, Agents | Expose 57 integrations and 754 actions, and explain that Agents support MCP |
| API productization | 2026 Current official website | API page | Public Completion, Embedding, Agent, Knowledge Folder API and model prices |
Langdock is a continuously iterative cloud platform, and the official version number of traditional software is not disclosed. Therefore, this article uses a "milestone version" to record its evolution: from its establishment in Berlin in 2023, to the current platform-based form for enterprise AI, to the expansion of developer capabilities in Integrations, MCP, and APIs. If the official Changelog subsequently provides a precise version number, the official Changelog should be used as the reference for updating.
At the model level, Langdock does not focus on self-developed basic models, but emphasizes model independence and unified access. The API page lists model suppliers such as OpenAI, Anthropic, Mistral, and Google, and gives examples of token prices for models such as GPT, Claude, and Gemini. The advantage of this route is that you can switch models based on cost, performance, region and compliance requirements; the limitation is that enterprises still need to understand the differences in capabilities and supplier terms of different models.
Technical advantages
Enterprise Governance First: Langdock’s product design revolves around the governance issues of enterprise adoption of AI. The Security page discloses controls such as EU hosting, no model training, AES-256, TLS 1.2+, least-privilege access, independent penetration testing; for regulated or multinational enterprises, these are closer to the core of purchasing decisions than "quality per answer".
Model agnostic and unified API: By using a unified API to call multiple model suppliers, enterprises can reduce the risk of model lock-in. For RAG, internal applications and Agent scenarios, unified Completion, Embedding, Agent, and Knowledge Folder APIs can reduce repeated encapsulation between different model interfaces.
Permission mirroring and integrated actions: The official security page mentions mirrored access permissions, and the Integrations page emphasizes that agents and workflows can search documents, trigger actions, and update data. For enterprise internal knowledge, whether the source system permissions can be inherited is more critical than "whether more documents can be indexed", because permission mismatch will directly bring the risk of data leakage.
MCP and custom integration: The Agents FAQ explicitly supports MCP and turns MCP tools into Langdock actions; the Integrations FAQ also explains that custom logic can be created for any API-enabled tool through the built-in integration builder. This allows Langdock to use off-the-shelf connectors or incorporate them into enterprise-built systems.
How to use
| Entrance | Applicable objects | How to use |
|---|---|---|
| Web workspace | Ordinary employees, business teams | Complete daily Q&A, document analysis and sharing assistants through Chat, Agents, Prompt library, Projects and other portals |
| Agents | Department heads, operations, support, legal affairs, sales | Configure instructions, knowledge, actions and triggering methods to create reusable internal Agents |
| Workflows | Operations, data, and process automation teams | Connect AI processing, data reading and writing, system actions, and notifications through workflows |
| Integrations | IT, Business System Administrators | Connect to tools like Google Drive, Confluence, Slack, Jira, Salesforce, and more, or build custom integrations using MCP/API |
| API | Developers, platform team | Use bearer token to call Completion, Embedding, Agent, Knowledge Folder API |
| Enterprise deployment | IT, security, compliance teams | Choose from multi-tenant, single-tenant, own cloud or on-premises Kubernetes/Helm deployment options |
Typical online steps are: first use a 7-day free trial to verify the employee Chat and knowledge Q&A experience; then select a department scenario to configure Agent and knowledge folder; then connect to key systems and verify permissions mirroring OAuth, logs and approvals; and finally expand to Workflows, API or enterprise deployment. True acceptance metrics should not just look at fluent responses, but also whether permissions are consistent, references are reliable, costs are predictable, and employees are willing to continue using it.
Product Pricing
Langdock's pricing structure has been disclosed on the Pricing page, which is divided into Chat & Agents, Workflows Add-on and API Add-on. Chat & Agents are basic subscriptions; Workflows scale by workspace and run volume; APIs are billed by model tokens. The official website also states that you can save 20% by paying annually, and all prices do not include VAT.
The public price for Chat & Agents is EUR 25/user/month for Standard seat and EUR 99/user/month for Business Max seat; Enterprise is Custom, for 1,000+ users, and supports dedicated deployment possible. Workflows Starter is included with Chat & Agents with 2,500 workflow runs/month; Workflows Business shows 40k runs/month for EUR 539/workspace/month and 100k runs/month for EUR 1199/month. Public examples of the API page include GPT-5.5 input EUR 4.76/1M tokens, output EUR 28.58/1M tokens, Sonnet 4.6 input EUR 2.6/1M tokens, output EUR 12.99/1M tokens, etc.; since the model price will change, the real-time price on the official website should prevail.
Items that are not disclosed or need to be confirmed include: Enterprise exact pricing, dedicated deployment fees, private cloud/on-premises commercial terms, higher workflow step caps, higher API rate limits, SLA and professional services fees. Before purchasing, it is recommended that the supplier provide a written quotation based on the target number of seats, expected monthly message volume, workflow running volume and API token consumption.
Application scenarios
- Internal knowledge assistant: Insert policies, procedures, product documents, support manuals and project materials into knowledge folders, allowing employees to query them through Chat or Agent. It is suitable for HR, IT, legal, sales support and other departments.
- Department-level repetitive task Agent: For example, scenarios displayed in official templates such as GDPR Privacy-Policy Auditor, IT Help Desk, Outlook Email Manager, BigQuery Analyst, Google Analytics Web Insights, etc. can save fixed instructions, knowledge, and tool actions as shared Agents.
- Cross-system workflow automation: Read forms, create tasks, send emails, update CRM, organize meeting minutes or generate reports with Workflows and Integrations, suitable for operations, finance, project management and customer success teams.
- Enterprise unified model gateway: The development team uses the Langdock API to uniformly call multi-model embedding, agent and knowledge folder, avoiding the need for each application to connect to the model supplier separately.
- Regulated Organization AI Promotion: Enterprises requiring EU hosting, GDPR, SOC 2 Type II, ISO 27001, SSO/SCIM/SAML, permission mirroring and deployment options can use Langdock as a more controllable AI workspace than public chat tools.
Applicable people
- Enterprise IT and Security Team: Teams that need unified management models, permissions, data resident SSO, auditing and deployment solutions are the core decision-making group of Langdock.
- Business Unit Leader: Teams that need to standardize repetitive work into Agents or Workflows, such as sales, customer service, operations, finance, legal, human resources and data analysis teams.
- Developers and Platform Teams: Teams that need unified model API, RAG, agent management and internal application access capabilities can use APIs to reduce the complexity of multi-model docking.
- Organizations with Larger Staffs: When an enterprise already has hundreds or thousands of employees who need to use AI securely, Langdock's governance, seats, and deployment capabilities are more likely to be of value.
The boundaries of unsuitability are also clear: when individual users only want to chat or write at a low cost, Langdock's corporate governance capabilities will appear to be biased; small teams without internal knowledge bases, system integration, and permission management requirements may not be able to use most of its platform capabilities for the time being.
Summary and Outlook
Langdock's core competency is to put several difficult points in enterprise AI adoption into the same platform: employee-friendly Chat, reusable Agents, internal knowledge, workflow, system integration, unified API and enterprise security compliance. Compared with a single AI chat tool, it is more suitable as an organizational-level AI workspace; compared with products that are only API gateways, it emphasizes a front-end experience that business users can directly use.
Points that currently require continuous observation include: whether the official Changelog discloses a more detailed version rhythm, whether the MCP and custom integration ecosystem can continue to expand, whether Enterprise deployment and pricing are transparent enough, and the impact of price changes of different models on the total cost of ownership. For enterprise buyers, the safest route is to start with a departmental knowledge assistant or high-frequency process agent pilot to verify permissions, references, usage and costs, and then expand to all-employee Chat, Workflows and APIs.
Related tools: 
CrewAI, 
LangChain
Version Info
- Langdock current public platform ：The Langdock platform currently displayed on the official website includes Chat, Agents, Workflows, Integrations, APIs, enterprise security and the ability to be deployed to EU, single tenant, own cloud or local environment; the continuous iterative cloud service does not disclose the traditional semantic version number.
- Langdock company established ：The official Press page disclosed that Langdock was established in Berlin in 2023. The product is positioned as an integrated platform to help enterprises safely promote generative AI; this article serves as a historical milestone record of cloud services.
User Reviews