Questions, answered plainly.
- What is ConceptSeek?
- ConceptSeek is a research instrument for long-form material. You build libraries from videos, playlists, podcasts, PDFs, and documents, then trace any concept through them. Results are exact passages from your own sources, with timestamps and links back to where they came from.
- How is ConceptSeek different from ChatGPT or Claude?
- ChatGPT and Claude generate text responses, which means they can hallucinate or paraphrase inaccurately. ConceptSeek does not generate answers. It retrieves verbatim passages from the sources you have added, including the source title and exact timestamp. Every result is directly citable, and the thinking stays yours.
- How is ConceptSeek different from NotebookLM?
- NotebookLM uses your sources as context for AI-generated responses. ConceptSeek returns the actual text from your sources, with no generation and no paraphrasing. This matters when you need to cite, because ConceptSeek results keep the original wording and source context. See a full comparison →
- What sources does ConceptSeek support?
- YouTube videos, entire playlists, channels, podcasts, PDFs, and text notes. Videos and podcasts are transcribed automatically on ingest; PDFs and notes are parsed directly. More source types are on the roadmap.
- What is Threads?
- Threads is ConceptSeek's relationship view. Every indexed moment from your libraries sits on a ring, and arcs between them show where your sources converge on a claim, diverge in contradiction, complement each other, or pose questions one another answers. When confidence is low, the tool says so.
- Does search use keyword matching or semantic search?
- Both. ConceptSeek runs a two-step pipeline: PostgreSQL full-text search for lexical precision, then vector similarity search using embeddings for conceptual relevance. Results from both stages are merged and ranked. You can search for an exact phrase or for the underlying idea.
- Can I search across multiple libraries at once?
- Yes. You select one or more libraries as the scope. All sources within those libraries are searched at once, and results are ranked by relevance regardless of which source they came from. Cross-library search is explicit on every query so you always know what you're looking inside.
- What is a library?
- A library is a named collection of sources you group together: one per project, topic, or research question. Search always runs within the libraries you select, so you control the scope.
- Who is ConceptSeek for?
- Researchers, journalists, students, analysts, and writers who work with large volumes of long-form material and need to locate specific ideas, arguments, and quotes with verifiable sources. It is built for people who need to cite, not just summarise.
- Is there a free plan?
- No. ConceptSeek is paid from day one, so it can stay independent and built for the people who actually use it. Every plan comes with a 7-day money-back guarantee on your first month. A public demo library (no signup required) is on the roadmap. See plan details on the pricing page.
- What happens to the content I add?
- Content you add is stored in your private account. It is not shared with other users, not used to train third-party models, and not accessible outside your account. Transcripts and embeddings are stored in a dedicated database tied to your user ID.
- Where can I see pricing?
- Current plans and tiers are on the pricing page.
Still have a question? We answer email personally.
Email support