{"id":16920,"date":"2026-06-05T16:25:17","date_gmt":"2026-06-05T14:25:17","guid":{"rendered":"https:\/\/www.canopsis.fr\/canopsis-26-04-when-ai-enters-open-source-hypervision\/"},"modified":"2026-06-10T09:33:45","modified_gmt":"2026-06-10T07:33:45","slug":"canopsis-26-04-when-ai-enters-open-source-hypervision","status":"publish","type":"post","link":"https:\/\/www.canopsis.fr\/en\/canopsis-26-04-when-ai-enters-open-source-hypervision\/","title":{"rendered":"Canopsis 26.04: When AI enters open source hypervision!"},"content":{"rendered":"\n

Canopsis 26.04 is now available<\/strong>, and it’s a milestone release! After 15 years of development and continuous evolution, our French open source hypervision solution<\/a> has reached a new milestone: the integration of artificial intelligence at the heart of the platform. <\/p>\n\n

Here’s what’s new for your teams \ud83d\ude80<\/p>\n\n

<\/div>\n\n

Introducing AI to Canopsis 26.04<\/strong><\/h2>\n\n

An AI assistant to create your filters in natural language<\/h3>\n\n
\n
\n
\n
\"Article<\/figure>\n<\/figure>\n<\/div>\n\n\n\n
<\/div>\n\n\n\n
\n

This is the flagship new feature of Canopsis 26.04: a Chatbot<\/a> is now integrated into the interface<\/strong>. It can be accessed from any modal containing a pattern<\/a> editor. <\/p>\n\n\n\n

More concretely, what does this change? <\/p>\n\n\n\n

Previously, creating or modifying a complex filter meant mastering the JSON<\/a> format. With Canopsis 26.04, that’s a thing of the past! You describe what you want in natural language and the wizard understands, suggests, corrects and generates the corresponding filter. It knows which fields are available and the constraints specific to each type of rule. <\/p>\n\n\n\n

The wizard is based on Google’s Gemini templates, configurable from: Administration \u2192 Custom objects \u2192 LLMs. <\/p>\n\n\n\n

This is the first step. The Capensis team has already announced that it is working on compatibility with other LLMs, including internally hosted models. This means that organizations subject to security or sovereignty constraints will also be able to benefit. <\/p>\n\n\n\n

This is a major step forward… But it’s only the beginning! Visit the Canopsis roadmap<\/a> to discover the next updates for the AI assistant. <\/p>\n<\/div>\n<\/div>\n\n

Anomaly detection also arrives in Canopsis 26.04<\/h3>\n\n

Alongside the AI assistant, Canopsis 26.04 also introduces an event stream<\/strong>anomaly detection<\/a> service <\/strong>. It is currently in alpha phase. <\/p>\n\n

Based on Isolation Forest<\/a> models, this service analyzes the volume of events received per connector, and reports abnormal behavior. Too many events, too few: both cases are detected automatically. <\/p>\n\n

Monitoring is managed from the Administration menu in Canopsis 26.04. You activate or deactivate connector-by-connector monitoring. The status of each monitored connector is then displayed in a new, dedicated Healthcheck<\/a> tab. <\/p>\n\n

Three states are possible: \u2705 OK, \u274c KO, \u2754 Insufficient data.<\/p>\n\n

It’s an experimental feature: it’s going to evolve quickly, so you might as well try it out now! As with the Chatbot feature, check the Canopsis roadmap<\/a> regularly for the latest updates. <\/p>\n\n

<\/div>\n\n
\n
<\/div>\n\n\n\n
\n
\n
\"Cycles<\/figure>\n<\/figure>\n<\/div>\n\n\n\n
<\/div>\n<\/div>\n\n
<\/div>\n\n

New features designed with teams in mind<\/strong><\/h2>\n\n

Beyond AI, Canopsis 26.04 also brings a number of much-awaited innovations to our hypervision solution community!<\/p>\n\n

A new alarm status: Unknown<\/h3>\n\n

When a parent entity fails, child entities may now appear with status Unknown<\/a>. This indicates that their problem is probably a consequence of the parent and not an autonomous failure. <\/p>\n\n

The result: fewer unnecessary tickets and a clearer picture of the situation.<\/p>\n\n

<\/div>\n\n
\n
<\/div>\n\n\n\n
\n
\n
\"Unknown<\/figure>\n<\/figure>\n<\/div>\n\n\n\n
<\/div>\n<\/div>\n\n
<\/div>\n\n

Comment templates<\/h3>\n\n
\n
\n

In Canopsis 26.04, drivers can now use <\/strong>comment templates<\/a> to structure and standardize their feedback on alarms<\/strong>.<\/p>\n\n\n\n

Composed of structured fields defined as mandatory or optional (time spent, problems encountered, notes…), the templates can be customized from the administration.<\/p>\n\n\n\n

To sum up, standardizing the comments entered enables :<\/p>\n\n\n\n