Product / Connectors
Your company knowledge comes from your systems.
COBACK reads from the systems you already use, from public registers and from your website. You do not have to upload anything.
- Google Drive
- SharePoint
- Slack
- Gmail
- Confluence
- Notion
- Box
- Archer
- PostgreSQL
- MySQL
- MongoDB
- Snowflake
- Substack
- RSS
- Mailchimp
- Langdock (via MCP)
- Your website
- Public registers (GLEIF)
Once connected, the digital twin stays current.
Every night COBACK reads what changed in the folders, sites and channels you shared, so the twin is up to date before the day's new rules are checked.
Every day, in this order
- 01
Connectors read
What changed in your systems since the last run.
- 02
The twin updates
New facts about your company reach the digital twin.
- 03
New rules arrive
From the official journals and authorities COBACK checks daily.
- 04
Relevance is checked
Each new rule against the current state of your company.
SharePoint
Through Microsoft Graph, COBACK reads the document library of the SharePoint site you name. Access is delegated and read only. Every night, changed files replace the old ones.
Each document keeps its link back to SharePoint and the name of the person who last edited it.
COBACK in your AI assistant
Your AI assistant connects to your workspace through COBACK's MCP interface. With an API key you create in COBACK, it can look up screening results, laws and tasks. A read-only key cannot change anything.
This works with any assistant that supports MCP, such as Langdock, Claude or your own agents.
All connectors
For every connector you decide what COBACK may read.
Documents and knowledge
SharePointOne site's document library
Google DriveOne folder of Docs, Sheets and files
ConfluenceOne space's pages
- NotionShared pages and databases
- BoxOne folder
Communication
SlackChannels the bot was invited to
GmailOne mailbox, read only
GRC
ArcherGRC records, tasks can be logged back to Archer
Databases
PostgreSQLOne table or view, read only
MySQLOne table or view, read only
MongoDBOne collection, read only
- SnowflakeOne table or view, read only
Newsletters and feeds
- RSSNew entries of an RSS or Atom feed
- SubstackThe posts of one publication
MailchimpCampaigns sent through an audience
Public sources
- Your websiteImprint, locations, products, certificates and linked reports
- Public registersLegal form, seat, register number and group parent via GLEIF
AI assistants and API
- LangdockAsks COBACK through the MCP interface
- Other assistantsAny assistant that supports MCP, with an API key
- REST APIScreening results, laws and tasks for your own systems
Connect, choose, stay current
Read access01Connect
You sign in with the provider and give COBACK read access.
- Scope
02Choose
You choose a folder, a site, a space, channels or a table.
- Every night
03Stay current
Every night COBACK reads what changed, and only what you chose.
What connectors may do
- Read onlyConnectors do not write to your systems. The one exception is Archer, where COBACK can log tasks if you want it to.
- Only the scope you chooseOne folder, one site, one space, the channels the bot was invited to, or one table.
- Encrypted credentialsCOBACK stores sign-in data encrypted with AES-256-GCM, or not at all.
- Your websiteThe crawler honours robots.txt, identifies itself as COBACK and stays on your domain.
- Public registersCOBACK asks GLEIF for legal form, seat, register number and group parent. A missing entry is never taken as proof that an entity does not exist.
Where COBACK stores and processes data is on the platform page.
More on securityQuestions about connectors
How often do connectors sync?
Every night. Connectors run before the daily check of new rules, so the assessments see the latest state of your company.
What does the Slack connector read?
The history of the channels you invite the COBACK bot to, threads included. It does not read direct messages.
Is there a connector for SAP, Teams or OneDrive?
Not yet. The overview above shows what COBACK can connect today.
Our system is missing. What now?
Talk to us. Many systems can be reached through a database, a feed or your website.