Ancher: turn saved articles, videos and notes into working knowledge
What Ancher is for
Ancher is a personal knowledge tool built around saving material you encounter and using it later. Its site describes capturing articles, videos, social posts, podcasts, PDFs and screenshots, then finding relevant information from a rough description instead of remembering the original location.
The product's focus is the movement from reading and watching to creating something from those sources. It offers automatically structured notes with key points and video keyframes, along with questions answered across the material in your saved library. Those are publisher-described capabilities, rather than results from Franklin's own account or testing.
From a source library to a draft
Ancher says newly saved material is connected with earlier sources, allowing related themes, contradictions and patterns to surface. Users can also place related notes into a collection and give that collection to the agent as the context for a particular job.
The stated outputs include memos, newsletters, research briefs, decks and posts. The collection step is useful to understand: it lets the user choose a narrower source set for a piece of work instead of asking the agent to draw indiscriminately from everything saved. The website also lists a browser-based app, mobile access and a browser extension for capture.
Start with a small collection
The free plan includes 100 credits per month, with paid plans offering larger allowances. The captured page does not establish how far those credits will stretch for every mix of saving and generation, so check the current allowance against your own workflow.
Franklin's suggested trial is to save a few sources on one topic, ask a question whose answer you can check, and create a short brief from that collection. Compare the draft with the originals before expanding the library. This tests retrieval and source use directly, rather than assuming that a larger collection automatically produces more reliable writing.