gnosis-air 0.1.5

The release of gnosis-air version 0.1.5 marks a significant advancement in the field of AI agent conversation analysis. This command-line interface (CLI) tool allows users to ingest and analyze conversation logs from AI agents, specifically those powered by Claude Code and Codex. By facilitating intent classification, friction analysis, and the creation of a cross-session knowledge graph, gnosis-air provides creators and technologists with a robust framework for understanding and improving AI interactions.

Features of gnosis-air

The gnosis-air CLI is designed to streamline the process of tracking AI conversations through a structured three-stage lifecycle: discovered, imported, and analyzed. This lifecycle ensures that the resource-intensive steps involving large language models (LLMs) — such as intent classification and pain-point detection — are executed only once. The results of these analyses are then stored for future reference, allowing users to revisit insights without incurring additional computational costs.

To utilize gnosis-air, users must have Python 3.12 or higher installed, along with an Anthropic API key or a local Ollama setup with the qwen2.5:3b model. The package is accessible via the Python Package Index (PyPI) and can be installed globally using the command `pip install gnosis-air`. This installation process also provisions three external CLIs that gnosis-air relies upon: jbcontext, chatter, and embark-call-graph, ensuring a smooth setup experience.

The CLI is designed with user convenience in mind. It allows for persistent session storage through a JetBrains AI token or an Anthropic API key, which can be set as an environment variable. This means that users can maintain their session state without needing to log in repeatedly. Additionally, the tool is built to be self-analyzing, automatically ingesting, classifying, and analyzing new conversations as they are recorded.

Technical Specifications and Usage

gnosis-air operates by creating a local SQLite database to store conversation logs. The database path can be customized, but the default location is `~/.local/share/gnosis/gnosis.sqlite3`. Users can interact with the CLI to perform various tasks, such as discovering new conversations, classifying intents, and analyzing pain points within the conversations.

The CLI's architecture allows for incremental processing of conversations. Each step in the analysis pipeline is designed to persist its output, meaning that if a user reruns an analysis on the same conversation, the system will only process new data, thereby saving time and computational resources. This is particularly beneficial for users working on large projects where conversations can be numerous and complex.

Furthermore, gnosis-air includes a mechanism for semantic search over both the codebase and the knowledge extracted from previous conversations. This feature enhances the user’s ability to navigate through extensive data, providing quick access to relevant tasks and insights. The CLI outputs results in a JSON format, making it easy to parse and integrate into other applications or workflows.

Why it matters

The introduction of gnosis-air 0.1.5 is a noteworthy development for creators and technologists involved in AI and machine learning. By providing a dedicated tool for analyzing AI agent conversations, it addresses a critical need for understanding user interactions with AI systems. The ability to classify intents and identify friction points can help developers refine their AI models, leading to improved user experiences.

Moreover, the self-analyzing capabilities of gnosis-air reduce the overhead typically associated with running LLMs, allowing teams to focus on insights rather than computational logistics. This efficiency is crucial as the demand for AI-driven solutions continues to grow across various industries.

In summary, gnosis-air 0.1.5 not only enhances the analytical capabilities available to AI developers but also sets a precedent for future tools in the domain. As AI continues to evolve, tools like gnosis-air will play an essential role in bridging the gap between technology and user experience, ultimately leading to more effective and user-friendly AI systems.