AI Session Summary [BETA]¶
Important
AI Behavioral Analysis is fully based on an internal model and does not send any data outside of Fudo Enterprise. In contrast, AI Session Summary relies on an external AI model, which requires sending session data to an external provider for processing.
Note
The AI-powered summaries for SSH sessions feature is available in Fudo Enterprise 5.6 as a beta version. You are encouraged to explore and try out this functionality to see how it can support your workflows.
Fudo Enterprise allows generating AI-powered summaries of selected SSH sessions. This feature helps administrators quickly understand session context without watching full recordings.
Note
- Session Summary is currently available only for SSH sessions.
- Summaries are generated on demand; automatic summarization is not yet supported.
- Session Summary requires prior configuration of an AI provider in the system settings.
Configuring AI Summary
Before using this feature, ensure proper configuration of AI model settings:
- Go to > > .
- Choose a provider (e.g., OpenAI or Anthropic), select a model, and paste a valid API key obtained from the provider’s platform.
- Alternatively, select the
CUSTOM
provider to connect a self-hosted model using the Ollama framework. In this case: -Model
must match the model’s name available on the Ollama server. -URL
must point to the server’s endpoint (e.g.,http://<ip>:<port>/api/chat
). - API key can remain empty.
Important
We recommend using larger models like OpenAI’s gpt-4.1
or Anthropic’s Sonnet
for optimal summary quality. Smaller models may reduce costs but often produce less accurate summaries.

Generating AI Summary
To generate a summary for one or more sessions:
- Select > .
- Locate the desired session(s).
- Click next to a session or select multiple sessions using checkboxes and click from the top menu.
- Wait for the summary to be created.

Interpreting Summary Results
Once the summary is generated, a corresponding status appears in the AI Summary column of the session list. Possible values include:
- No risk found – the session was analyzed and no potential security risks were identified.
- <number> potential risks found – the analysis detected one or more risk indicators within the session.
- Failed – the summary could not be generated due to an error (e.g., model misconfiguration or API issue).
To review the generated summary, click the status value in the row of the respective session.

Related topics: