Flexnote

Research becomes hard to reuse when the quotation is in a PDF, the example is halfway through a video and the reason you saved either is in a separate note. Flexnote puts those pieces on a visual board, where a question, its evidence and a possible explanation can sit together. For a video essay or course outline, the useful outcome is being able to return to the exact source while the script is changing.

For a video essay, one board could hold a question, three competing explanations and the source material supporting each. Keep direct quotations visibly separate from your own summary. The canvas can make relationships easier to inspect, but a convincing arrangement of cards is not evidence by itself. You still need to open the original and check context.

Build a source trail you can still follow later

The provider describes PDF highlights becoming reference cards and timestamped notes linking back to videos. That is a useful workflow for course creators combining written references with demonstrations. Place a claim, its source and the intended chapter near one another, then draft your explanation only after the evidence is clear.

The homepage also advertises Markdown and bidirectional linking, along with import integrations for existing knowledge collections. Before moving an established archive, test a small representative set. Check whether links, annotations, attachments and exports retain enough context to be usable outside the canvas. We have not tested import fidelity or migration of a large collection.

The AI layer should be evaluated as retrieval over your material, not as proof that the material is true. Ask a question whose answer you already know is in one imported document, then check whether you can identify the supporting passage. Keep your own interpretation distinct from a quotation. A knowledge assistant can make a source easier to retrieve while still summarizing it incompletely.

Compare Kuse when your priority is an AI-assisted research workspace rather than a local-first visual board. ↗

The free notebook and the AI allowance are different limits

The observed free tier provides three whiteboards and 100 cards, basic canvas features, local storage, basic export and one device. That is a reasonable way to assess the working style. It should not be presented as unlimited research storage or full multi-device synchronization. A creator who stores one card per quotation may reach the card ceiling sooner than expected.

The pricing page listed Pro at $29 per quarter or $49 per year, with a separate lifetime option. It also displayed promotional and education prices. We are not presenting those offers as Picktool-negotiated discounts or assuming their duration. Check the exact purchase terms and renewal basis on the current provider page.

Card limits are easy to underestimate. A research board with seventy sources and two cards per source would contain 140 cards before you add the outline or task notes. That is a planning illustration, not a claimed Flexnote project we built, but it shows why the 100-card free ceiling matters. Smart Index points are a different quota; buying indexing capacity does not redefine the free notebook limit.

Paid plans listed 3,000 Smart Index points, and additional one-time packs were advertised. The explanation connects indexing to note text and text in images for semantic search and AI retrieval. Larger documents and complex images can consume the allowance differently, so a provider example estimating a number of notes should not be treated as a guaranteed capacity.

The pricing FAQ says keyword search and AI chat continue after the indexing allowance is exhausted, but long documents and PDFs may give chat less material to draw on. That makes the budget question more specific than asking whether AI is included. Evaluate whether the documents you actually research become retrievable in the way you need.

What local-first means when you enable AI

The current privacy policy makes the AI boundary more specific: Smart Index, OCR and AI requests can send the text or image needed for the request to Alibaba Cloud DashScope. The policy says Flexnote does not train its own models on your notes, while DashScope processes requests under its own terms. That is a narrower statement than saying no provider ever receives or uses the material. Optional sync introduces a separate storage service again.

The official desktop availability is Windows and macOS. We did not verify a native phone app, test offline AI inference or measure the advertised large-canvas performance. If mobile capture and effortless team co-editing are essential, confirm those requirements directly rather than assuming them from the phrase knowledge base.

Our Gemini Notebook guide discusses another approach to connecting sources and AI-assisted research. ↗

Try one small research brief: a PDF, a useful video timestamp, several connected notes and an exported outline. Ask whether the board helps you find the exact source when revising a script a week later. That is a stronger evaluation than filling a canvas with attractive cards. Flexnote is most compelling when spatial organization supports an existing research habit, not when maintaining the system becomes another project.