fluctor

Your data stays yours.

Hosted in the EU or entirely on your own servers, with access set area by area and every change logged.

Read the DPA

Six commitments, in full.

EU hosting

Documents, memory and every AI call stay in the EU, and a call that fails moves to a second EU route, never outside it. Only operational metadata may be processed elsewhere, under standard contractual clauses.

Your own servers

fluctor runs entirely on your own hardware: documents, memory and AI models stay on your premises. Only usage figures, status codes and your team’s logins go to fluctor’s cloud, never a document, an email or an answer.

GDPR

fluctor processes your data as your processor, only on your instructions, under a data processing agreement based on Article 28 GDPR. When the contract ends, the data is deleted completely. The exact terms are in the DPA.

No training

Your data never trains a model. It is processed to run fluctor for you and for nothing else.

Every source shown

Every answer and every signal shows the document it was read from, and one click opens it. If your data holds no match, the assistant says so instead of guessing.

Activity log

Every change is logged: who, what, when and the outcome, for people and for every agent run. The activity log shows the last 365 days and can be exported as CSV or JSON; a write-once copy is kept for seven years and cannot be deleted, not even by an admin.

Where your data lives.

Two ways to run fluctor: in EU data centres, or on servers you own.

In the EU

fluctor runs in European data centres. Each company’s documents, memory and answers are stored under its own company ID.

  • Documents and memory are stored in the EU.
  • AI calls run in EU regions only.
  • Access tokens of email and drive connections are stored encrypted.
  • When the contract ends, the data is deleted completely.

On your own servers

fluctor installs on your hardware and runs inside your network. Documents, memory and AI models stay on your premises.

  • It reads databases, internal APIs and file servers inside your network.
  • Only usage figures, status codes and your team’s logins go to fluctor’s cloud.
  • Apart from that sync, the agent’s web search is the one call that leaves your network, and it needs access an admin adds in Settings.

Who sees what.

Companies kept apart

Every record is stored under your company’s ID, and that ID comes from your login, never from a request. Nothing crosses from one company to another.

Areas for each team

An admin creates areas such as Purchasing and Finance and grants them person by person. Search, signals, cases, dashboards and the assistant show only what a person’s areas allow, and an agent reads with the areas of the person it runs for.

Managing is not reading

An admin manages areas without opening them. To read one, the admin grants it to themselves, and that shows in the member list and in the activity log.

Support only when invited

The fluctor support team can work in your company for up to two hours, and only after one of your admins starts the session. It ends by itself, you can end it any time, and every action in it is logged as support.

The data processing agreement.

Published in full, under Article 28 GDPR. These are the clauses IT and legal ask about first.

Open the DPA
1
ScopeApplies wherever fluctor processes personal data on your behalf.
6
Security measuresMeasures that fit the risk, as Article 32 GDPR requires. An updated description on request.
7
Sub-processorsListed on request. Every change is announced, with ten business days to object.
9
AuditsOnce every twelve months, with thirty days’ notice.
10
Data breachesReported to you without undue delay.
11
Return and deletionDeleted completely when the contract ends. A readable copy on request.

Questions IT asks first.

Each answer names the place to read it.

  • Your documents, memory and AI processing stay in the EU. Service providers may process operational metadata on servers in the United States, safeguarded by standard contractual clauses under Article 46 GDPR.

    Source: Privacy policy, section 4.3

  • Yes. It installs on your hardware and runs inside your network, with documents, memory and AI models on your premises. Apart from the usage figures and your team’s logins that go to fluctor’s cloud, the one call that leaves your network is the agent’s web search, and it works only after an admin adds access in Settings.

    Source: Settings, AI & processing, Models

  • Everyone sees the general area. Every other area is granted to a person by an admin, and search, signals, dashboards and the assistant leave out what the person may not see.

    Source: Settings, People, Areas

  • Only during a support session that one of your admins starts. It lasts up to two hours, ends by itself, and every action in it appears in the activity log as support.

    Source: Settings, Organization, General

  • No. An agent’s first send waits for you, and Approve this run releases only that run. It can email only the recipients you named; with an empty list it cannot send email at all.

    Source: An agent’s page, Does without asking

  • The activity log in the app shows the last 365 days. Admins can filter it by person and event, and download the same view as CSV or JSON. A write-once copy is kept for seven years and cannot be deleted, not even by an admin.

    Source: Settings, Organization, Activity log

  • When the contract ends, the data is deleted completely. The exact terms are in the DPA. On written request you also get a readable copy of the data.

    Source: Data processing agreement, clause 11

See fluctor on your own documents.

Bring your IT team. We cover hosting, access and the activity log on a real setup, not on slides.