Fluid Topics
Fluid Topics is an enterprise product knowledge platform that unifies scattered technical documentation into a single hub, then delivers it to portals, chatbots, in-product help and AI agents through semantic search and built-in RAG services.
What is Fluid Topics?
Fluid Topics calls itself a Product Knowledge Platform, a category it positions as the successor to the Content Delivery Platform it launched in 2014. The publisher is ANTIDOT SAS, a Lyon company founded in 1999 that has since consolidated everything under the Fluid Topics brand.
The problem it addresses is fragmentation. In organisations selling complex products, technical content is scattered across authoring tools, PDF archives, SharePoint sites and knowledge bases, and none of it is shaped for the AI systems companies now want to build on top. Fluid Topics ingests that material through ready-made connectors, restructures and enriches it with consistent metadata and taxonomies, and holds it in a unified knowledge hub. Delivery is headless: the same hub feeds documentation portals, chatbots, in-product help, helpdesk tools and AI agents at once, with permissions managed from a single point. DITA is supported in full, including specialisations, customisations, cross-references and keys.
Search combines keyword and semantic retrieval, personalises results by role and behaviour, turns any metadata into a filter and reindexes in real time. On top of that sit native RAG services: embedding computation, vector search, prompt management and pre-built LLM connectors, all of which run internally so that content never leaves the platform. Generative features cover question answering, summarisation, code assistance and support-agent reply suggestions, always with citations. An Agentic AI capability exposes a production-ready MCP server for connecting autonomous agents. Translation is generated on demand rather than maintained in advance.
A WYSIWYG Portal Designer lets non-technical staff build homepages, search pages and reader pages, and an analytics suite reports on searches, content gaps, navigation, bookmarks, AI usage and the cost savings from self-service. Offline mode and social features such as commenting, sharing and bookmarking round out the portal. The platform targets software, manufacturing and high-tech electronics companies, and names Teradata, Darwinbox, Hexagon, Johnson Controls, Liebherr and Siemens among its customers.
What it does
- Ingest and continuously sync product content from CCMS, LMS, SharePoint and more than 100 sources and formats
- Publish a branded, permission-controlled documentation portal without writing code
- Answer user questions with combined keyword and semantic search across every content type
- Ground chatbots and virtual assistants in trusted product knowledge, with citations back to the source
- Connect autonomous AI agents to the knowledge hub through a production-ready MCP server
- Translate content on the fly into the reader's language instead of maintaining pre-translated sets
- Measure content usage, knowledge gaps and self-service deflection through a built-in analytics suite
When to use Fluid Topics / When not to
A quick filter to help you decide if Fluid Topics is the right fit.
When to use Fluid Topics
- Documentation teams at software vendors and manufacturers juggling thousands of DITA topics and legacy PDFs across several publishing channels
- CIOs and IT leaders who need to make existing product content AI-ready without rebuilding the underlying infrastructure
- Customer support organisations chasing ticket deflection through a self-service portal backed by a grounded chatbot
- Field service teams whose technicians need contextual, permission-filtered documentation on mobile, including offline
- Product and engineering groups that must keep published documentation aligned with every release of a complex product
When not to use Fluid Topics
- Individuals and small teams: this is a contract-based enterprise platform with no self-service sign-up and no published price
- Anyone looking for an authoring tool, since Fluid Topics delivers content but does not create structured content the way a CCMS does
- Buyers who need to evaluate legal terms up front, because no privacy policy, terms of service or DPA is published on the site
- Teams that want to trial software before talking to sales, as neither a free plan nor a free trial is advertised
- Users expecting a native mobile application, the product being served through a responsive web portal only
How to use Fluid Topics
A typical end-to-end flow, from setup to results.
- Start from the website: there is no self-service sign-up, so the entry point is a demo request or a contact form
- Optionally size the opportunity first with the ROI calculator or the AI readiness assessment offered on the site
- Browse the demo library after registering, to see the portal, search and AI capabilities in context
- Work through scoping and commercial terms with the vendor, since pricing is quoted rather than published
- Connect your existing content sources using the ready-made connectors, keeping your current authoring processes intact
- Let the platform restructure and enrich the ingested content, applying metadata and taxonomies for search and AI
- Design the documentation portal in the WYSIWYG Portal Designer, controlling layout, branding and landing pages without a developer
- Configure access management so each audience, role and group sees only the content it is cleared for
- Publish straight to production in real time, with no staging environment or FTP workflow to maintain
- Monitor processing, usage and content gaps from the admin interface and the analytics suite, and refine from there
Pros & Cons
Pros
- ISO 27001:2022 certified for Fluid Topics, with the certificate published openly as a downloadable PDF
- Embeddings and the vector database run internally, so content is never exposed to an external provider
- Customers choose their own LLM, with pre-built connectors and the freedom to switch if answer quality disappoints
- Data residency is negotiable, and customer data can be confined to a single location such as the United States on request
- Extensive ready-made connectors across more than 100 sources and formats, with no change to existing authoring processes
- Published customer outcomes are concrete: 65% support deflection at Darwinbox, five million portal visits at Teradata, two weeks saved per release at Hexagon
- Twenty-six years of prior work in semantic search and natural language processing behind the product
Cons
- No privacy policy and no terms of service anywhere on the site, an unusual gap for an enterprise SaaS vendor
- No data processing agreement published, and none advertised as available on request
- No subprocessor list and no data retention schedule are disclosed
- The vendor never states whether customer content is used to train models, in either direction
- Pricing is entirely private, so budgeting is impossible before engaging with sales
- Neither a free plan nor a free trial is advertised, and there is no native mobile application
- No direct support email is published: help goes through a portal and contact forms only
Pricing & Plans
Fluid Topics does not publish any pricing. There is no pricing page on the site, no plan grid and no indicative amount, and no free plan or free trial is advertised. The commercial model is quotation-based: prospects are invited to request a demo or contact the vendor, after which terms are agreed individually. An ROI calculator is offered on the site, but it estimates value rather than disclosing list prices. Prospective buyers should therefore expect an enterprise sales cycle and budget accordingly.
Data, GDPR & hosting
A consolidated view of how Fluid Topics handles your data.
GDPR overview
Fluid Topics is operated by ANTIDOT SAS, established in Lyon, France, so the controller sits inside the European Union and no Article 27 representative is required. The security page explicitly claims compliance with GDPR, CCPA and other applicable regulations, and the platform is ISO 27001:2022 certified, with the certificate published as a downloadable PDF. A Data Protection Officer is named on the legal page, Mr Stéphane Loesel, reachable at dpo@antidot.net. Access and rectification rights are described under Article 34 of the French Data Protection Act of 6 January 1978. The gap is documentary rather than declarative: no privacy policy, no terms of service, no data processing agreement, no subprocessor list and no retention schedule are published anywhere on the site.
Who owns the data?
No terms of service or privacy policy is published on the site, so ownership of customer content is never stated in writing. What the vendor does document is custody rather than title: Fluid Topics is a multi-tenant application in which customer databases are kept separate and access is gated by application logic. Fluid Topics staff may reach a customer's settings, documentation and analytics for support purposes only with that customer's approval. Separately, the legal page states that prospect data collected through forms and the newsletter is used by Antidot alone for surveys, analysis and marketing, and remains in its exclusive possession. Contractual ownership terms must therefore be settled directly with the vendor.
Reuse rights
The site publishes no terms of service, so there is no public statement on whether an end user may reuse content without asking permission. What is documented concerns processing rather than reuse rights. Embeddings and the vector database run inside the platform, and the vendor states that content is never leaked to external sources. Access rights follow the content into retrieval-augmented generation, so a user only ever receives passages they are already cleared to see. Each customer picks their own large language model, and the vendor points customers to that provider's own privacy policy for anything beyond its perimeter. Audit logging and governance controls are available. Notably, the site never states whether customer content is used to train models.
Data retention & training
Hosting summary
Fluid Topics operates data centres in the United States, Europe and Asia, and states that it works with customers to meet specific requirements, including confining customer data to a single location such as the United States only. The application is multi-tenant, with customer databases kept separate and file storage protected by application logic that grants access to the relevant customer alone. Testing and staging environments are separated from production, and no production data is used in development or test. The network is segmented into multiple security zones, with databases, application servers, source code and CI/CD tooling held in the most trusted ones. All communication is encrypted with AES-256 at rest and TLS 1.3 in transit, falling back to TLS 1.2 only where older devices require it. One caveat on jurisdiction: the public website resolves to a Cloudflare anycast address, which reflects the content delivery network and says nothing about where customer data actually resides.
Things to keep in mind
Risks and trade-offs to weigh before adopting Fluid Topics.
- The contractual framework cannot be reviewed before contacting sales: no terms of service and no privacy policy are online
- No data processing agreement is published even though the platform processes customer content, so one must be demanded during negotiation
- No subprocessor list is disclosed, leaving it unclear which third parties touch the data
- No retention period is published, so how long content and personal data are kept is unknown
- The vendor never states its position on training models with customer data, which is precisely the question buyers now ask first
- Because the language model is chosen by the customer, part of the applicable privacy posture depends on that third party rather than on Fluid Topics
- Over-reliance on AI-generated answers can erode the habit of reading source documentation; citations are provided, but only if readers actually follow them
Setup & Integrations
Technical difficulty
Moderate, and deliberately front-loaded onto the vendor. A full portal is advertised as launchable in a few hours, the Portal Designer needs no technical skills, ready-made connectors cover more than 100 sources without changing existing authoring processes, and publishing goes straight to production with no staging environment. An onboarding team assesses and replicates existing connectors during a migration. That said, this remains an enterprise deployment: single sign-on, role-based access control, permission modelling and integration with the wider information system all require IT involvement, and custom integrations through the APIs assume developers on the customer side.
Deployment
Integrations
Behind Fluid Topics
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Fluid Topics.
Frequently asked questions
What exactly is Fluid Topics?
Who is behind Fluid Topics?
How much does it cost?
Does it replace a CCMS?
Is there an API?
Where is the data hosted?
Does my content get sent to an external LLM?
Is the platform certified?
Is there a privacy policy or a DPA?
Which competitors does the site name?
Should you pick Fluid Topics?
Fluid Topics is a mature, technically serious answer to a problem that only gets worse as products grow more complex: product knowledge scattered across authoring tools, file archives and knowledge bases, none of it ready for the AI systems companies now want to build. Twenty-six years of work in semantic search sit behind it, and the customer outcomes the vendor publishes are specific enough to be checkable, from 65% support deflection at Darwinbox to five million portal visits at Teradata.
Its strongest differentiator is architectural. Embedding computation and vector search run inside the platform rather than at a third party, access rights follow content all the way into retrieval, and customers keep the choice of language model. Combined with ISO 27001:2022 certification, regular third-party penetration testing and negotiable data residency, that makes a credible case for organisations where governance is the binding constraint.
The reservation is documentary rather than technical, and it is a real one. The site publishes no privacy policy, no terms of service, no data processing agreement, no subprocessor list and no retention schedule, and it never states whether customer content is used to train models. For an enterprise vendor that is otherwise thorough about security, the silence is conspicuous, and every one of those points will have to be settled in contract negotiation instead of being read beforehand. Pricing is equally private, so no budget can be framed without engaging sales.
The verdict follows from the positioning. This is not a tool an individual evaluates on a Friday afternoon; it is an enterprise platform for organisations that already carry a heavy documentation estate and a real governance requirement. For that audience it deserves a serious look, provided the legal paperwork is demanded early.
- Choosing a selection results in a full page refresh.
- Opens in a new window.