01
Verifiability over fluency
A confident paragraph with no source is worse than no answer at all. Citations are part of the output, never an afterthought.
✦AI Document Intelligence
Document Intelligence
for Serious Work
Chronicler turns the files you already trust into an interactive knowledge base. Ask in plain language and get answers with citations, document attribution, and a confidence indicator.
PDFWordExcelPowerPointPlain textScanned pages (OCR)PDFWordExcelPowerPointPlain textScanned pages (OCR)
001who we are
Chronicler is a desktop-native knowledge assistant that turns hundreds of scattered documents into one library you can actually question, with every answer traced back to the page it came from.
5+document formats
100%answers with citations
2run modes: desktop & self-hosted
5+document formats
100%answers with citations
2run modes: desktop & self-hosted
002why chronicler
Why Chronicler?
Most tools hand you a confident paragraph and hide where it came from. Chronicler does the opposite: the evidence arrives with the answer.
02
Privacy over convenience
Desktop-native by default and self-hostable when it matters. Your library is not training data.
03
Replaceable over locked-in
Every model and store in the pipeline can be swapped, so the product outlives any single vendor.
003capabilities
What Chronicler Does
01
Conversational Document Search
Ask questions in plain language across your entire library instead of opening files one at a time.
- ✓ Natural-language queries
- ✓ Whole library or a single collection
- ✓ Follow-up questions keep their context
✦
02
Multi-Format Ingestion
PDF, Word, Excel, PowerPoint, and plain text are parsed, chunked, and indexed the moment you import them.
- ✓ Native parsers per format
- ✓ OCR for scanned pages
- ✓ Re-indexing when files change
⌁
03
Collections
Group related documents and scope a question to exactly the material that matters.
- ✓ Named document groups
- ✓ Chat scoped to one collection
- ✓ Reusable across conversations
✦
04
Cited, Verifiable Answers
Every response carries its citations, the documents behind it, and a confidence indicator.
- ✓ Inline source citations
- ✓ Jump back to the original passage
- ✓ Confidence shown, not implied
⌁
05
Teams & Workspaces
Seats, roles, and admin controls for organisations running Chronicler together.
- ✓ Role-based access control
- ✓ Seat assignment and reclaim
- ✓ Audit log and GDPR erasure
✦
✦?
Not sure where to start?
Import one folder and ask one question. The answer will tell you whether the rest of your library belongs here.
Your Documents. Yours Alone.◆ Source-cited answers◆ Role-based access◆ Self-hostable stack◆ GDPR erasure & audit log
004the product
See It Working
Three moments from the desktop app: bringing documents in, asking across them, and checking the passage an answer came from.
A library, not a folder
Import a directory once. PDFs, Word, Excel, PowerPoint, and scanned pages go through the same pipeline and land in collections you control.
Ask in plain language
Put a real question to the whole corpus. The answer comes back with document attribution and a confidence indicator attached, not as a floating paragraph.
Open the source
Every claim carries its citation. Click through to the passage itself and check it before the answer changes a decision.
005process
How It Works
From a folder of files to an answer you can defend: five steps, no manual searching.
Import your documents
Point Chronicler at the PDFs, Word files, spreadsheets, and decks you already work from.
Automatic indexing
Text is extracted, OCR'd where needed, chunked, and embedded into a searchable index.
Ask in plain language
Query your whole library, or narrow the question down to a single collection.
Retrieve and rerank
The pipeline pulls candidate passages, reranks them, and keeps only what is actually relevant.
Answer with sources
You get the answer, the citations behind it, and a confidence indicator you can act on.
006use cases
Who It's For
01ASK → RETRIEVE → CITE
01
Researchers drowning in papers
Hundreds of PDFs, one question at a time. Ask across the whole corpus and get back the passage itself, not a summary you cannot verify.
See how it works ↗Citedevery claim traceable
Corpusone question, all papers
007the stack
Open and Swappable
ChroniclerRAG engine
OpOpenAIOlOllamaFaFastAPINeNext.jsOpOpenAIOlOllamaFaFastAPINeNext.js
ReReactElElectronPoPostgreSQLpgpgvectorReReactElElectronPoPostgreSQLpgpgvector
SQSQLAlchemyAlAlembicDoDockerUvUvicornSQSQLAlchemyAlAlembicDoDockerUvUvicorn
Every component in the pipeline (embeddings, vector store, reranker, and language model) can be replaced. Run against OpenAI, or keep the whole thing local with Ollama.
008pricing
Built for Teams
of Any Size
Chronicler is licensed per seat. Bring your own OpenAI key or run it locally: there is no per-query bill to watch.
Essential
For a small team running a real pilot
$15/ seat / month
What's included:- ✓ 2–5 seats
- ✓ 750 indexed files
- ✓ Full pipeline with citations
- ✓ Desktop app
Most Popular
02Team
For growing teams
$25/ seat / month
What's included:- ✓ Everything in Essential
- ✓ 5–25 seats
- ✓ Unlimited indexed files
- ✓ Collections & workspaces
✦
Need the whole organisation?
Talk to usUncapped seats, audit log, GDPR erasure, and a dedicated contact.
009team
The Team Behind Chronicler
010FAQs
Common Questions
PDF, Word (.docx), Excel (.xlsx), PowerPoint (.pptx), and plain text. Scanned pages go through OCR first, so image-only PDFs stay searchable.
Chronicler is a desktop application, so your library is indexed on the machine or server you run it on. In the default configuration the retrieved passages are sent to your configured language model provider to compose the answer. Point it at a local Ollama model if you would rather nothing leaves the machine.
Every answer arrives with citations, document attribution, and a confidence indicator, so you can open the source passage and check it before you rely on it.
No. Install the desktop app, import a folder, and ask a question. Self-hosting the backend with Docker is optional, and it is documented step by step.
Yes. The stack is FastAPI, PostgreSQL with pgvector, and a Next.js frontend orchestrated with Docker Compose, so it deploys to your own VPS or internal network.
Still have a question?Contact us ↗