Frequently asked questions.

Everything that matters. Your documents, the text Chauncy extracts from them, the search index, the embeddings, and your annotations all live on your local disk. Chauncy doesn't require an account, doesn't phone home, and doesn't track usage.

Two things, both only when you initiate them:

1. URL ingestion — when you paste a URL, Chauncy fetches the page from the web. The request goes out, the page comes back, Chauncy indexes it. After that, no further network activity on that document.
2. AI synthesis — if you turn AI on, only the excerpts Chauncy selects to answer your specific question are sent to your chosen provider, and only when you ask. Your full collection, index, and annotations stay local. If you use Ollama or a local OpenAI-compatible endpoint, nothing leaves the machine at all.

Yes for search, indexing, and AI synthesis using a local model such as Ollama or a local OpenAI-compatible endpoint. The internet is needed only for URL ingestion, cloud AI providers, and voice input, which uses your system's speech recognition and may require network.

No. Buy the installer through Stripe, install it, run it. No account, no login, no profile.

No. Chauncy's semantic search works without any AI provider. AI synthesis is optional — turn it on when you want it, leave it off when you don't.

Whatever your chosen provider charges. Chauncy uses your own API key with Anthropic or OpenAI; you pay the provider directly for what you use. Chauncy doesn't mark up provider costs, doesn't route your requests through our servers, and doesn't see your prompts or responses. If you run a local model with Ollama or a local OpenAI-compatible endpoint, there's no provider cost at all — just the hardware you already own.

Chauncy doesn't bundle a local AI model. To use one, you install model-serving software on your own machine and download a model that fits your hardware. Ollama is a common choice. Chauncy connects to the running model through its local API. The setup is well documented by those projects and typically takes a short session to complete. If you'd rather skip the setup, cloud providers and search-only mode are available without any of this.

PDF, Word (.docx), PowerPoint (.pptx), Markdown, HTML/HTM, plain text, and email files (.eml, .mbox). Web pages can be ingested by pasting a URL.

Different tools, different jobs.

• Note-taking and PKM tools ask you to build a knowledge graph by hand — tagging, backlinking, naming things consistently. Chauncy doesn't ask for that work. It infers the structure from the meaning of what you've written.
• Desktop search tools match keywords. Chauncy matches meaning. A search can find a relevant document even when it shares no keywords with your query.
• Read-it-later services save articles to read. Chauncy makes the things you've saved anywhere searchable by meaning.

You can use Chauncy alongside any of these. Chauncy is the layer that lets you find and synthesize across what you've collected, however you collected it.

Not at launch. Chauncy supports one collection per installation. You can ingest from many folders, mailboxes, and URLs, but everything lands in the same searchable library.

A perpetual license to install and use Chauncy on computers you own or control, plus all updates within the current major version. New major versions may carry separate pricing; existing customers are notified in advance.

Windows 10 or 11, 64-bit. 4 GB RAM minimum, 8 GB recommended for larger collections. Roughly 1 GB free disk for the application and runtime, plus space for your indexed content as it grows. No GPU required.

Not yet. macOS support is in development. There's a waitlist in the footer.

Email support@chauncy.dev with the subject "Refund Request" within 14 days of your purchase. No explanation required. Refunds are processed within 3-5 business days through Stripe back to your original payment method.