mAItion
Turn document chaos into one trusted AI-powered brain
About mAItion
mAItion is an open source knowledge management platform built around a constraint most enterprise AI search tools treat as optional: every answer must be verifiable. Organisational knowledge is scattered across Confluence, Google Drive, Slack, Jira, email and half a dozen SaaS tools, and the usual response is a chatbot that confidently summarises all of it with no way to check whether the summary is true. mAItion returns exact citations and full source context with every response, so a claim can be traced back to the document it came from in a single click.
The project is built by WikiTeq, a team with deep MediaWiki heritage, and is used by organisations including Wikipedia, the United States Space Force, Novartis, NASA, NATS and AMS. It is released under the MIT licence with an actively maintained Python codebase on GitHub, so it can be self-hosted rather than adopted as a closed SaaS dependency, which matters a great deal to the regulated and government users on that list. Topics on the repository place it squarely in the RAG and knowledge-graph space, with Open WebUI integration and a connector architecture. Because the source is open, the retrieval behaviour can be audited rather than taken on trust.
Connect your existing systems through the connector library, which covers Confluence, Google Drive, Slack, Jira, Microsoft 365, Teams, SharePoint, Notion, Salesforce, HubSpot, Zoho, Gong, email and MediaWiki. mAItion indexes across all of them so a single search reaches every system rather than requiring you to remember which tool a document lives in. Ask a question through the chat interface and the answer arrives with exact citations and the surrounding source context attached, so verification is a click rather than a research task. Background AI agents continuously refresh and enrich the knowledge base so entries do not silently go stale as the underlying documents change. Self-host from the repository, or engage WikiTeq for a managed deployment.
- •One Search Across Every System - Confluence, Google Drive, Slack, Jira, Microsoft 365, Teams, SharePoint, Notion, Salesforce, HubSpot, Zoho, Gong, email and MediaWiki unified into a single knowledge hub
- •Citation-Backed Answers - Every response carries exact citations and full source context, so any claim can be verified against its origin document in one click
- •Continuous Knowledge Refresh - AI agents keep the knowledge base current and enriched rather than letting indexed content drift out of date
- •MIT Licensed and Self-Hostable - Fully open source, so it can run inside your own infrastructure with no data leaving your control and no vendor lock-in
- •Auditable Retrieval - Because the code is open, how documents are retrieved and ranked can be inspected rather than assumed
- •Open WebUI Integration - Works with the Open WebUI ecosystem rather than forcing a proprietary front end
- •Automated Workflows - Actions can be triggered from what the knowledge base surfaces, not just questions answered
- •Enterprise Deployment Track Record - In use at organisations including Wikipedia, NASA, the US Space Force, Novartis, NATS and AMS
Fits engineering and operations teams whose documentation has sprawled across too many systems to search reliably, and support teams who need an answer they can quote to a customer with a source attached. It is a particularly strong fit for regulated, government and research organisations where a cited, self-hosted, auditable system is a hard requirement rather than a preference, which is reflected in its existing user base. Teams evaluating closed enterprise AI search products should weigh the MIT licence, since self-hosting removes both the recurring cost and the question of where the documents are being processed.
Pricing
$10/mo
- Standard$10/mo
- EnterpriseContact sales
From the vendor pricing page, 2026-09-19
