- ThurWiki
- Connectors
Your systems, on the page. Without a marketplace.
A company wiki describes what sits in other systems: customers in the accounts, issues in the ticket system, configurations in the repository. A copy typed out by hand is out of date the moment the source changes. ThurWiki fetches the content from the source onto the page; one click brings it up to date, and every refresh appears in the history. And in the other direction, your systems reach into the wiki through its interface.
This is how it looks on the page.
Customer from bexio
Master data, outstanding invoices, one invoice with its line items. Fetched from bexio instead of typed out; one click brings the latest state.
Project from bexio
The project page tells what is being built; the recorded hours and the quotation it fetches for itself.
Issues from Jira
Assignee, status, priority. Whoever reads the page need not open Jira and needs no account there; who may see it is decided by the space.
File from the repository
The configuration out of Git, with the latest changes underneath.
41 connectors in thirteen groups.
A connector is access to another system, set up in the wiki administration. The keywords underneath show which macros or functions it brings into the wiki.
Commercial
- Contact
- Invoice
- Quotation
- Invoice list
- Project
- Recorded hours
- Customer (WHMCS)
- Services
- Tickets
- Single ticket
- Record (Abacus)
- List
- History
ERP
- Customer or supplier
- Invoice (ERP)
- Invoice list (ERP)
- Order or quotation
- Product
Source code
- Issue list (Git)
- Single issue
- File from the repository
- Latest changes
- Pull and merge requests (not Gogs)
Ticket systems
- Issue
- Issue list
- Project
Jira
- Issue (Jira)
- Issue list with JQL
- Project (Jira)
Software builds
- Build status (Jenkins)
- Code quality (SonarQube)
Operations
- Host status
- Open problems
- Graph, drawn in the wiki
- VM or container
- Cluster overview
File storage
- File from Seafile
- Folder from Seafile
- File from Nextcloud
- Folder from Nextcloud
CRM
- Customer
- Person
- Opportunity
- Case
- Notes
Groupware
- Insert a contact
- Appointment
- Insert appointments
Messaging
- Insert an e-mail
- Notification by e-mail
- by Telegram
- by WhatsApp
Artificial intelligence
- AI assistant
- Voice input
- Transcripts of voice notes
Search and images
- Image search
- Generate an image
- Web search in the assistant
Three principles.
- Read-only. From bexio, WHMCS, Abacus, Odoo, Flectra, Jira, Git, ticket systems, Jenkins, SonarQube, Zabbix, Proxmox, Nextcloud, Seafile, CRM, calendar, address book and mailbox the wiki only fetches, it never writes back; access there needs read rights only. The messaging connections send notifications, and the AI connections send questions and, for a transcript, audio recordings to the provider you chose.
- Rights from the wiki. Whoever may see the page sees the content, even without an account in the source system. The space decides.
- Secrets stay secret. Stored encrypted, never shown again, neither in the interface nor through the API.
And in the other direction: your systems reach into the wiki.
Anything the interface can do, a program can do, because the interface uses the very same API. 36 endpoints, 257 actions, documented. Monitoring creates fault pages, the ticket system fetches the instruction, the intranet shows the newest pages.
- An API key as a service account. It gets rights like a person, can belong to groups, and appears in the log under its own name. As narrow as you cut it.
- Markdown in and out. Create and read pages as Markdown; upload attachments; spaces, pages, versions, comments, tasks, search and users are all addressable.
- A key is not a master key. 64 characters, shown once, after that stored only as a checksum; it cannot sign in and cannot switch wiki.
WIKI=https://wiki.your-company.com
# Read a page as Markdown
curl -s -H "Authorization: Bearer $TW_KEY" \
"$WIKI/api/pages.php?action=get&id=412&format=markdown"
# Create a page from monitoring
curl -s -H "Authorization: Bearer $TW_KEY" \
-d action=create -d space_id=12 \
-d title="Outage 11 September" \
--data-urlencode content_markdown@report.md \
"$WIKI/api/pages.php"
# Full-text search
curl -s -H "Authorization: Bearer $TW_KEY" \
"$WIKI/api/pages.php?action=search&q=emergency+power"One endpoint per file, actions as parameters, answers as JSON. The documentation ships with the wiki.
Questions about the connectors
What is the difference from Confluence marketplace apps?
Three things. The connectors are part of the product, with no price per user and per app. They are maintained together with the wiki, not by a third party that disappears at the next update. And the data sources are connected read-only: access to bexio, Jira or Git needs read rights only, so a mistake in the wiki can do no harm there.
Who sees the content from a connected system?
Whoever may see the wiki page. The rights come from the wiki, not from the source. The engineer sees the invoice on the page about the job without needing a bexio account. Whoever uses such a macro decides, through the space, who sees it.
What does it take to set up a connector?
An access key or a user with read rights in the source system, and the address of the service. The wiki administrator enters both once in the administration and tests the connection with a button. Secrets are stored encrypted and never shown again.
My system is missing. What now?
Talk to us. The connectors share a common frame; anything with a documented interface can be connected. We discuss what is missing with you, and much of it we build into the product, with no surcharge per app.
A wiki with integration for bexio, WHMCS, Abacus, Jira, GitHub, GitLab, Bitbucket, Gogs, Mantis, Redmine, Request Tracker, Zammad, YouTrack, Jenkins, SonarQube, Zabbix, Proxmox VE, Nextcloud, Seafile, SuiteCRM, vtiger, CalDAV calendars, CardDAV address books, e-mail mailboxes, Telegram and WhatsApp: ThurWiki shows invoices, quotations, projects, tickets, issues, pull requests, build results, hosts and problems from monitoring, virtual machines, files from the repository and from file storage, customers, opportunities and appointments as building blocks on the wiki page, fetched from the source and refreshed with one click. Read-only access to the data sources, rights from the wiki, no marketplace apps, plus AI connections to Claude, ChatGPT, Gemini, DeepSeek, LiteLLM, Ollama and Whisper.
Which source belongs on your pages?
We show you the connectors in the demo wiki, on request with test access to your own bexio, Jira or CRM.