Add to Chrome
Chrome extension · No card to start

Your team's wiki,
finally answers back.An AI reading assistant that chats with your PDFs, docs, and wikis — and cites every answer.

Sidenote sits in your browser and answers questions about Confluence pages, runbooks, PRDs — and any PDF, doc, or article you can open. Every answer cites the source passage.

Add to Chrome — freeWatch the citation demo

7-day Pro trial · No card required · Free tier forever

getsidenote.app/security-complianceSidenote
SidenoteDocumentationData handling

How Sidenote handles your data

Updated 14 May · sidenote-team · public

Two retention modes per document, your choice. Store keeps the document indexed for the lifetime of your account so re-opening doesn't re-pay for indexing. Discard purges everything within 24 hours of upload via an hourly background sweep.

All embeddings live in Supabase pgvector in our UK (eu-west-2) region, partitioned per user. Row-level security policies enforce per-account isolation — no two accounts can ever read each other's chunks.

sidenoteTrialStore
How Sidenote handles your data
Cached41 chunksStore
SummariseChatExplainGlossary
ThreadsNew
How long does Sidenote keep my docs?
Store keeps the document while your account is open 1; Discard purges it within 24 hours of upload 2. Embeddings live in the UK 3, isolated per account 4.
Ask a question…
Article · Sidenote docsv1.12.0
Row-level securityCitations, no fabricationSemantic retrieval
What it reads

Built for internal wikis first.
Works on whatever else you open.

Confluence pages, Notion docs, runbooks, PRDs, PDFs, papers, articles, even Slack canvases — if it loads in your browser, Sidenote can read it.

Confluence
SharePoint
Google Docs
PowerPoint decks
arXiv papers
Linear specs
Substack
Wikipedia
Reddit threads
Web pages
Notion
OneDrive
Word documents
PDFs
GitHub READMEs
Medium articles
Slack canvases
Stack Overflow
Hacker News
Compared honestly

Sidenote vs the notebook-style tools.

If you live in your wiki and your browser, you want answers where you’re already reading — not in a separate “research workspace”.

CapabilitySidenoteNotebook tools
Lives in your browser, on the page you’re reading Yes No
Works on Confluence, Notion, Google Docs, Slack, SharePoint & OneDrive today Yes No
Citations scroll and highlight the source passage Yes No
Open the page once, ask questions later (caching) Yes Partial
Ingest PDFs, articles, web pages on demand Yes Yes
Audio overviews No Yes

We’re not a generalist research workspace. We do one thing: read the page you’re on, with citations. See how Sidenote compares to NotebookLM →

The signature feature

Click a citation. Watch the source light up.

Every answer carries inline markers. Click one and Sidenote scrolls the page and pulses the exact passage. When the model bookends a topic by quoting first and last sentence, each cited sentence lights up on its own — try 7. If the document doesn’t say it, Sidenote doesn’t either.

Try clicking the chips
How long do you keep my documents?
Store mode keeps the document while your account is open 1; Discard mode purges within 24 hours of upload 2.
Where are embeddings stored, and is my data isolated?
Embeddings live in Supabase pgvector in our UK region 3, isolated per account by row-level security 4.
Can the model invent things the document doesn’t say?
No. Every answer is checked server-side and unsupported claims are dropped before the response leaves the API 5.
Will I burn my quota re-opening a document I’ve already ingested?
No — re-opening from cache is free and doesn’t count against the monthly cap 6.
Are my chats used to train AI models?
No — both of our AI sub-processors operate under no-training defaults on the API tiers we use 7.
Sidenote · Documentation · Data handling

Sidenote offers two retention modes per document. Store mode keeps the document indexed for the lifetime of your account, or until you delete it. Discard mode purges everything within 24 hours of upload via an hourly background sweep.

All embeddings are stored in Supabase pgvector in our UK (eu-west-2) region, partitioned per user. Row-level security policies enforce per-account isolation at the database layer — no two accounts can ever read each other’s chunks.

Chat answers stream over server-sent events; every claim is validated server-side against the retrieved passages before it leaves the API. Unsupported claims are dropped automatically, and the model’s verbatim quotes are rectified to the chunk’s exact text so citation chips land on the precise passage.

Documents are content-hashed and shared across users, so two people ingesting the same source only pay once for indexing. Re-opening a document from cache is free and doesn’t count against the monthly cap, and runs on a separate, more generous hourly bucket.

Sidenote relies on two AI sub-processors. Anthropic provides Claude for chat, summaries, and explanations; Voyage AI provides the embedding engine. Anthropic retains prompts and responses for 30 days under their standard data policy, while Voyage AI does not retain embedding inputs. Both operate under no-training defaults on the API tiers we use.

For internal wikis

Your Confluence isn’t a knowledge base.
It’s a graveyard with search.

Every fast-growing team has the same dirty secret. Twelve thousand pages, half of them stale, the same five answers re-asked in standup every week. Sidenote doesn’t replace your wiki — it teaches it to answer. See the full pitch for Confluence →

Before

Open 14 Confluence tabs, ctrl-F through six runbooks.

After

Ask one question, get one answer, jump to the source line.

Before

New engineer pings the team channel: “where do we deploy hotfixes?”

After

New engineer asks the wiki. The wiki answers. The team is at lunch.

Before

Three pages disagree on retention. Which is current?

After

Sidenote reads all three in a collection, surfaces the disagreement, and cites each so you can verify.

How it works

From cluttered tab to confident answer in three moves.

Step 01

Open any page

A Confluence runbook, a PDF you just downloaded, an arXiv paper, a long Substack — Sidenote sits in the side-panel of every tab.

Step 02

Ingest once

We extract, chunk, and embed the page in seconds. Re-opening the same doc is free — re-ingest from cache, no cap hit.

Step 03

Ask with citations

Summarise, chat, explain, build a glossary. Every claim links to the exact passage — click and the document scrolls and pulses.

Use cases

Picks up where ctrl-F gives up.

Same product, different roles. The thing in common: you’re tired of re-reading the same documents to find one paragraph.

01

Engineering

Runbooks, RFCs, on-call playbooks, postmortems. Ask the wiki what we did last time — get the link to the doc that says so.

02

Product

PRDs, research transcripts, competitive notes. Synthesise across twenty docs without losing the receipts.

03

Customer Success

Knowledge-base articles, escalation runbooks, prior tickets. Get the answer with the source citation, ready to paste.

04

Legal & Ops

Policy docs, vendor contracts, compliance evidence. Quote the exact clause; never guess at it.

05

Research

Long PDFs, arXiv papers, lit reviews. Pull supporting passages straight into your notes.

06

Everyone else

Substacks, Medium long-reads, that 14,000-word Stratechery post. Sidenote works on whatever loads in your browser.

See every source Sidenote reads, or compare the plans.

FAQ

Everything we’d ask if we were you.

NotebookLM is a research workspace you upload sources into. Sidenote lives on the page you're reading. We optimise for one thing: in-browser, with citations that scroll the source. We don't do audio overviews; we do “click this chip, see the paragraph”.
Yes. Confluence reads through the Chrome extension on the page you can already see — no integration to set up, no admin install, no SSO scopes to grant, and your wiki content goes through your own account, not a shared pipeline. Notion, Google Docs, Slack canvases, and SharePoint & OneDrive connect through each provider's official API with read-only access to the documents you choose to open, again through your own account. SharePoint and OneDrive sign in with Microsoft (delegated, read-only Graph scopes); on locked-down Microsoft 365 tenants a one-time admin approval may be needed.
No. Anthropic and Voyage AI run with no-training defaults on the API tier we use, and we never fine-tune models on user content. Your documents are used only to answer your own questions.
Web articles (Readability extraction) and browser-rendered PDFs work today, alongside dedicated integrations for Confluence, Notion, Google Docs, Slack canvases, and SharePoint & OneDrive. For everything else, drop a PDF, DOCX, PPTX, TXT, MD, MDX, or RTF file onto the side-panel and Sidenote ingests it the same way. Uploads chat headlessly — citation chips highlight passages in the upload's text rather than scrolling a tab you don't have.
Every answer is checked server-side against the retrieved passages before it leaves the API. If a claim has no source, the chip is dropped — you only ever see citations the model could actually back up. Click a chip and Sidenote scrolls the document to the exact passage and pulses it in amber. When the model bookends a multi-sentence topic by quoting first and last sentence, each cited sentence is highlighted on its own — the gap in the middle stays plain, so you see exactly what was used to ground the answer.
Chat runs on Anthropic Claude — Sonnet on Pro for fast multi-doc chat, Opus on Pro+ and Pro Max for sharper reasoning across long, dense docs. Chunk summaries and glossaries use Claude Haiku. Embeddings come from Voyage AI's voyage-3. We pick the model so you don't have to.
Data lives in Supabase's eu-west-2 (UK) region, encrypted at rest, with row-level security so no two accounts can read each other's chunks. Store mode keeps the document for the lifetime of your account, or until you delete it. Discard mode purges the document — including its chunks and any orphaned shared content — within 24 hours of upload via an hourly background sweep.
Sign up with email, click the magic link, and the 7-day Pro trial starts — no card required. Trial caps are 3 documents, 10 explanations, 3 glossary builds, and 10 chat turns across the week. We only ask for a card when you convert. After that, one click in the dashboard opens the Stripe Customer Portal — cancel, change plans, or update your payment method any time. You keep Pro access until the end of the current billing period.
Ready when you are

Stop digging. Start asking.

Add Sidenote to Chrome, open any page in your wiki, and ask it the question you’ve been Slacking the team about.

Add to Chrome — freeSee the citation demo

7-day Pro trial · No card required · Free tier forever