Inventory

Product Introduction

1. Definition: Inventory is a local-first, private desktop application that functions as a unified semantic search index for AI-powered coding and development conversations. It is a technical productivity tool that operates as a read-only aggregator and search engine for local chat histories.
2. Core Value Proposition: It exists to solve the problem of lost context and knowledge fragmentation across multiple AI coding agents (like Cursor, Claude Code, Zed, and Codex) by creating a single, searchable timeline of every conversation, enabling developers to instantly retrieve past solutions, code snippets, and discussions without relying on cloud services or manual organization.

Main Features

1. Unified Cross-Agent Indexing: Inventory automatically reads and snapshots the local data stores of supported AI coding tools (IDEs and agents) to build a consolidated, private search index. It works by parsing local SQLite databases or log files in a read-only manner, ensuring no interference with the original tools. The core technology involves on-device file system monitoring and data normalization.
2. Hybrid Semantic & Keyword Search: Every conversation is processed using on-device embedding models (like those from the sentence-transformers family) to create vector representations. Searches blend traditional keyword matching with semantic similarity search, allowing queries like "container stuck" to find relevant conversations about "pod stuck terminating." Results are clearly labeled to distinguish between keyword and semantic matches.
3. Quick Capture & Contextual Resume: The ⌘⇧N (Quick Capture) hotkey allows instant saving of a thought or question, which is immediately matched against the entire indexed history. For supported agents like Claude Code, the tool can reopen a terminal session with the full historical transcript attached, enabling the AI to resume work with full context, even if the project directory has changed.
4. Clipboard History Scratchpad: An optional system-wide ⌘⇧V hotkey activates a searchable, app-tagged history of everything copied to the clipboard. This feature is off by default for privacy, uses macOS accessibility APIs for global hotkey detection, and stores data locally with clear data collection disclosures.
5. Configurable History Retention & Privacy Controls: Users can set the index to retain conversations from the last 7, 30, 90, 365 days, or indefinitely. The interface displays the precise on-disk storage impact of each choice, providing transparent data trade-offs. All processing is done locally; no data is sent to the cloud.

Problems Solved

1. Pain Point: Knowledge Silos and Context Loss. Critical solutions, debugging steps, and configuration commands become trapped and forgotten within the isolated chat histories of different AI coding tools, leading to repetitive problem-solving and wasted time.
2. Target Audience: Software Engineers, DevOps Professionals, and Technical Leads who regularly use multiple AI-assisted development environments (Cursor, Zed, Claude Code) and need to maintain a persistent, searchable knowledge base of their technical work and decisions.
3. Use Cases: A developer debugging a Kubernetes error can search their hybrid index for "image pull backoff" to find the exact Helm configuration fix they discussed with Claude Code two weeks prior. A team lead can copy a condensed primer of a complex solution thread and paste it into a Slack thread or new chat session to hand off context seamlessly.

Unique Advantages

1. Differentiation: Unlike cloud-based note-taking apps (Notion, Obsidian Sync) or tool-specific history, Inventory requires no manual copy-pasting, no subscription, and no cloud account. It differs from system-wide search tools (Spotlight, Alfred) by performing deep, meaning-based search within structured conversation data from niche professional tools.
2. Key Innovation: Its read-first, local-only aggregation model is the key innovation. Instead of being another cloud service or a new writing surface, it passively unifies the proprietary, local data silos created by modern AI-native IDEs. The hybrid search (keyword + on-device semantic vectors) applied specifically to conversational developer history is a novel technical approach to the problem of personal knowledge retrieval.

Frequently Asked Questions (FAQ)

1. How does Inventory ensure my chat data remains private? Inventory operates entirely on your local machine. It reads chat histories in a read-only mode, creates its own local search index, and does not transmit any data to external servers. All processing, including semantic embedding, happens on-device.
2. Which AI coding tools and IDEs are compatible with Inventory? Inventory currently supports local conversation history from Cursor, Claude Code, Zed, and Codex. The platform is built to be extensible, with support for Windows and additional tools like Kiro and others planned for future updates.
3. What is the difference between keyword search and semantic search in Inventory? Keyword search matches the exact words in your query to words in your chat history. Semantic search uses AI models to understand the meaning of your query and find conversations with similar context, even if they don't share the same keywords (e.g., finding "pod stuck terminating" when you search for "container stuck").
4. Is Inventory a subscription service? No, Inventory is a one-time purchase license with a single fee of $19.99. This includes the current version and all future feature updates for life, with no recurring subscription or account required.
5. Can I control how much disk space Inventory uses? Yes. Inventory provides granular control, allowing you to index only the last 7, 30, 90, or 365 days of conversations, or all history. The settings menu clearly shows the estimated disk space usage for each option before you apply it.