Service Discontinuation Notice - Kompas AI

What

Kompas AI was an AI research product that provided in-depth research using agent-based methods such as multi-agent orchestration and context window management. Based on the notice, it appears to have been offered as a B2C service and positioned around helping individuals access deeper AI-assisted research workflows.

The service is being discontinued, with an official closure planned for December 2025. The source page does not provide detailed workflow, feature, or audience segmentation beyond the fact that it used multiple applied AI-agent techniques and is ending in favor of a new project focused on AI agent permission control and safety management.

Features

• In-depth research output: The product was described as delivering in-depth research, suggesting a focus on more comprehensive AI-generated analysis rather than lightweight answers.
• Multi-agent orchestration: Kompas AI used multi-agent techniques, which likely helped coordinate specialized AI tasks within a single research process.
• Context window management: The service applied context window management, a practical capability for handling larger or more structured information during research tasks.
• Applied AI-agent architecture: The notice references a range of AI agent–based technologies, indicating the product was built around orchestrated agent workflows rather than a single-model interface.
• Consumer-facing service model: The page explicitly identifies Kompas AI as a B2C service, which implies it was intended for end users rather than solely enterprise teams.

Helpful Tips

• Plan for product sunset risk: Since Kompas AI is scheduled to close in December 2025, teams or users should avoid depending on it for long-term operational workflows.
• Verify migration paths early: The notice mentions Peta as a new project, but it does not state that Peta replaces Kompas AI’s research functionality, so assumptions about continuity should be avoided.
• Assess the underlying pattern, not just the product: The most durable takeaway is the use of multi-agent orchestration and context handling for research, which can inform evaluation of alternative AI research tools.
• Separate research capability from governance capability: Kompas AI focused on research, while Peta is described around permission control and safety management, so buyers should treat these as different product categories unless future documentation states otherwise.
• Request documentation before adoption of similar tools: For products in this category, it is important to confirm source handling, workflow transparency, and data retention practices, especially when public product pages are sparse.