Summarizes, extracts, and rewrites content from files.
Airgap
FreeStarting price
0Popularity
About Airgap
Airgap is a local document AI tool designed for confidential files. It enables users to query PDFs, Markdown, and text files using local retrieval, local generation, and citations that link back to the original pages. All processing occurs on the user’s machine, ensuring documents never leave the local environment. The tool builds a private vector index for each vault, parses documents, retrieves relevant chunks, and runs a local language model to generate answers with precise citations. Airgap supports five built-in models or any GGUF model from a Hugging Face repository, and it operates offline once models are downloaded. It is intended for local research workflows where data residency, provenance, and precision are prioritized over cloud-based convenience. The application is available for macOS, Windows, and Linux, with vaults stored as portable files protected by the machine’s disk encryption.
Key features
- Local retrieval and generation of answers
- Citations linking answers to original document pages
- Support for PDF, Markdown, and text files
- Five built-in models plus custom GGUF model compatibility
- Multilingual retrieval with a 100-language embedding model
- Portable and shareable vaults stored locally
- Disk encryption protection for vaults at rest
- Optional connected model for hosted API access
Use cases
- Research workflows requiring data confidentiality and provenance
- Onboarding new team members with project-specific knowledge
- Querying technical documentation or code repositories locally
Pros
- Processes documents locally without cloud dependency
- Provides paragraph-level citations linking answers to source pages
- Supports multiple file formats including PDF, Markdown, and text
- Offers offline operation after model download
- Allows use of custom GGUF models from Hugging Face
Cons
- Requires local model downloads for offline operation
- Connected model feature is optional and off by default
- No cloud-based account or workspace integration
Frequently asked questions about Airgap
What is Airgap and what does it do?
Airgap is a local document AI tool that allows users to query PDFs, Markdown, and text files using local retrieval, generation, and citations. All processing occurs on the user’s machine, ensuring documents never leave the local environment.
Who is Airgap designed for?
Airgap is designed for local research workflows where data residency, provenance, and precision are prioritized over cloud-based convenience. It is suitable for individuals or teams handling confidential files.
How does Airgap ensure security and privacy?
Airgap operates with local custody of files, using disk encryption such as FileVault, BitLocker, or LUKS to protect data at rest. Documents are never uploaded, and all processing, including retrieval and generation, occurs offline once models are downloaded.
Can I use my own models with Airgap?
Yes, Airgap supports five built-in models and allows users to bring any GGUF model from a Hugging Face repository. Users can also point Airgap at a hosted API for connected model use, though this is optional and off by default.
What file formats does Airgap support?
Airgap supports PDFs, Markdown, and text files. It builds a private vector index for each vault, enabling efficient retrieval and generation based on the content of these formats.
How do I get started with Airgap?
Users can download Airgap for macOS, Windows, or Linux from the official website. After installation, they can create a vault, add documents, and begin querying locally without requiring an account or cloud workspace.
