Question 1

What is SeekFiles AI?

Accepted Answer

SeekFiles AI is a Retrieval-Augmented Generation (RAG) assistant that lets you upload files and ask questions about them. Every answer is grounded in your content and shows the exact source chunks it used.

Question 2

How is this different from just uploading a file to ChatGPT?

Accepted Answer

We index every chunk of your files into a private vector database with hybrid retrieval (semantic + BM25 keyword) so the assistant can find the right passage even in large libraries. You can scope assistants to specific files or folders, share public-facing assistants, and every answer cites the source — none of which a one-off chat upload gives you.

Question 3

Do I need a credit card to start?

Accepted Answer

No. The free tier gives you 50 welcome credits and up to 3 assistants. You only add a payment method if you upgrade to a paid plan or buy a top-up.

Question 4

How fast can I be up and running?

Accepted Answer

Most people upload their first file and ask a question in under two minutes. Larger uploads take longer to embed but you can keep working while indexing happens in the background.

Question 5

What file types are supported?

Accepted Answer

PDFs (including scanned ones via OCR), Word documents, plain text, Markdown, EPUBs, spreadsheets (XLSX, CSV), presentations (PPTX), and ZIP archives that contain any of the above. Up to 100 MB per file.

Question 6

Can it read scanned PDFs and images?

Accepted Answer

Yes. Our vision pipeline renders each PDF page to an image, OCRs the text, and captions any figures or charts. Scanned contracts, handwritten notes, and image-only PDFs become fully searchable.

Question 7

What languages are supported?

Accepted Answer

Cross-lingual retrieval works out of the box — ask in one language, retrieve from source files in another. The underlying multilingual embeddings cover most major languages.

Question 8

Is there a limit on how many files I can upload?

Accepted Answer

Yes — storage caps scale with your plan. Free starts at a smaller cap; Starter, Pro, and Business each lift the ceiling. Per-file size is 100 MB on every tier.

Question 9

How do citations work?

Accepted Answer

Every assistant message comes with citation chips showing the exact file and chunk used. You can click a chip to see the source passage and verify the answer is grounded in your content.

Question 10

Does it hallucinate?

Accepted Answer

We have a hallucination guard that detects when retrieved chunks aren't strong matches for your question and tells the model to refuse rather than fabricate an answer. Combined with a similarity floor on citations, it keeps weak grounding from sneaking through.

Question 11

Can I scope an assistant to specific files?

Accepted Answer

Yes. Each assistant has its own system prompt and an attached set of files or folders. An assistant only retrieves from its scope — useful for separating, say, your finance docs from your research papers.

Question 12

Which language model is used?

Accepted Answer

GPT-4o-mini by default for cost-efficiency, with GPT-4o available per-assistant when you want maximum quality. You pick the model when configuring each assistant.

Question 13

How does pricing work?

Accepted Answer

All plans are credit-based. Subscriptions give you a monthly credit refill plus an assistant allowance. Top-up packs add one-time credits that never expire. See the pricing page for current numbers.

Question 14

What is a credit equivalent to?

Accepted Answer

Credits track LLM usage — input tokens, output tokens, and embeddings. A typical short Q&A on GPT-4o-mini costs roughly 1 credit; longer answers or GPT-4o cost more. The dashboard shows the exact debit per message.

Question 15

Do credits roll over?

Accepted Answer

Yes. Each renewal adds your monthly allowance to your existing balance — unused subscription credits stay in your account and stack with any top-ups. Top-up credits also never expire on their own.

Question 16

Can I cancel anytime?

Accepted Answer

Yes. Billing is handled by Lemon Squeezy. Cancel from your account in one click — your plan stays active until the end of the period.

Question 17

Is there a mobile app?

Accepted Answer

Yes — native iOS and Android apps. Both work alongside the web dashboard, syncing your files, assistants, and conversation history.

Question 18

Can I share an assistant publicly?

Accepted Answer

Yes. On Starter and above you can mark an assistant as public and share it via link. Public chats use a configurable credit multiplier so you control your spend.

Question 19

Can I embed an assistant on my website?

Accepted Answer

Public assistants can be opened via a shareable URL today; an iframe-friendly embed is on the near-term roadmap.

Question 20

Who can see my files?

Accepted Answer

Only you and accounts you explicitly grant access to. Files and embeddings are scoped to your account and never used to train any third-party model.

Question 21

Where is data stored?

Accepted Answer

Files live in object storage with at-rest encryption. Embeddings are stored in a Postgres pgvector database. Both are within infrastructure that we control.

Question 22

Can I delete my data?

Accepted Answer

Yes. Deleting a file removes it from storage, deletes its chunks and embeddings, and detaches it from any assistants that referenced it.

Questions, answered.

Getting started

Files and content

Answers and accuracy

Pricing and credits

Mobile, web, and sharing

Privacy and security

The free tier is the fastest way to know if it fits.