Deepnote pricing explained — what Free, Team and Enterprise actually include

Deepnote pricing at a glance (read 1 August 2026)

• Free — "Free forever". Up to 3 editors, up to 5 projects, "Limited Deepnote AI", "Unlimited Basic machines with 5 GB RAM, 2 vCPU", and 7 day revision history. Viewers are unlimited on every plan.
• Team — $39 per editor per month billed yearly, shown with a -20% badge. The pricing page opens on the Yearly toggle. The monthly figure is $49, and it does not appear anywhere in the HTML the server sends — see the next section. Team adds unlimited editors and projects, "$39 worth of AI credits every month", "$280 worth of CPU every month", "$50 worth of GPU every month", "GPT-5.5 and Sonnet 4.6 access", 30 day revision history and a 24 hour inactivity period.
• Enterprise — "Custom", with a "Talk to us" CTA rather than a price. It is the only tier where "Bring your own LLM" is ticked, and it adds SSO and directory sync, audit logs, permission groups, single-tenancy, private Docker images and unlimited revision history.
• Team is the only tier with a self-serve trial: the CTA reads "Try 14 days for free".
• There is no student or education tier on the pricing page — the words "education", "student" and "academia" appear zero times in it. The Education plan exists, but only in the documentation: "Deepnote is completely free for students, teachers, and researchers", conditional on a school email address.

Why you will see $39 quoted everywhere and pay $49

Deepnote's pricing page has a Yearly / Monthly toggle and it opens on Yearly. The Yearly view shows $39 with a -20% badge and the label "per editor/month billed" followed by "yearly". That is a completely normal way to present annual pricing, and Deepnote labels it accurately.

What is unusual is where the monthly number lives. We downloaded the pricing page on 1 August 2026 and searched the raw HTML: the string "$49" appears zero times. The only dollar figures in the document Deepnote's server sends are $39 (twice), $50, $280 and $112. The monthly rate exists in the page's own JavaScript bundle, as the object price:{month:"$49",year:"$39"}, and is rendered into the DOM only after you click Monthly.

The practical consequence is small but real, and it is the sort of thing worth knowing before a budget conversation. Anything that reads the page without executing its JavaScript — a scraper, a price-comparison site, an AI assistant summarising the page — sees $39 and not $49. So the number that propagates outward is the annual one. If you are budgeting monthly for a five-person team, the figure is $49 x 5 = $245 a month, not $39 x 5 = $195. Committing annually is a genuine 20% saving; it is just a commitment.

Where Deepnote's own pages disagree with each other

This is the part a listing page will not have, because it requires reading Deepnote's pricing page and its documentation side by side on the same day. None of these are accusations of bad faith — every fast-moving product accumulates drift between its marketing page and its docs. But each one is a place where you could plan against the wrong number, so they are worth knowing.

• A plan called Pro is documented but not sold. Deepnote's own pricing FAQ says "Deepnote currently offers 3 plans: Free, Team, and Enterprise." Three separate documentation pages disagree: "Deepnote AI is available on Pro, Team and Enterprise plans", "Deepnote Agent is available on Pro, Team and Enterprise plans", and "Scheduling is available on Team, Pro and Enterprise plans". The word "Pro" appears zero times as a plan name on the pricing page. If you read the docs first, you may go looking for a tier that is not on sale — and, more importantly, you cannot tell from the docs whether Deepnote AI and the Agent are available on Team, because the sentence granting them lists a plan that no longer exists alongside it.
• Free compute is described as both unlimited and worth $112. The Free plan card says "Unlimited Basic machines with 5 GB RAM, 2 vCPU", and the comparison table two sections down ticks "Basic (5 GB RAM, 2 vCPU)" as Unlimited for Free. The same table then heads a row "Free compute ($112/month worth)". Deepnote's machine-hours documentation takes the first side: "Our goal is to provide an unlimited number of free machine hours for all workspaces." The tooltip on the $112 row explains the Basic-versus-Plus distinction and never mentions a cap or the figure. We cannot tell from Deepnote's published material whether the $112 is a real ceiling or a valuation of what the free hours would otherwise cost.
• The $112 row also does not reconcile with the Team card on the same page. That single row spans all three columns, yet the Team card separately promises "$280 worth of CPU every month" and "$50 worth of GPU every month". $112 is neither of those numbers nor their sum.
• The Basic machine spec is not on the hardware page. Deepnote's "selecting hardware" documentation contains no CPU machine specifications at all — it says only "More powerful machines are available on paid plans." The 5 GB / 2 vCPU figure for Basic appears in the documentation only on the Education FAQ, and the Team default of "unlimited hours on a 16GB, 4vCPU machine" appears only on the Projects concept page. Neither is where you would look for it.
• The listed GPUs are two generations old in the docs. The hardware page lists "4 vCPUs, 64 GB memory, K80 GPU" and "8 vCPUs, 64 GB memory, 16 GB V100 GPU", while the pricing page's row reads "GPU (60 GB RAM, 12 GB K80 GPU)" — 60 GB against 64 GB, and a V100 option the pricing page does not list. If GPU work is why you are buying, ask sales rather than trusting either page.

What the Free plan actually caps, in the order you will hit it

Deepnote's Free tier is more generous than most in this category on compute and less generous than most on AI, and the limit you hit first depends almost entirely on what you connect it to.

• The AI quota goes first, and it goes fast. Free gets 10 AI code completions per month, and 5 calls per month each for Auto AI, Generate code, Edit code and Explain code. That is a demonstration allowance, not a working one — a single afternoon of AI-assisted work exhausts it. The card's phrase "Limited Deepnote AI" is accurate. Deepnote's docs confirm the shape without numbers: "Each plan includes a monthly AI quota, and usage beyond the quota is billed as pay-as-you-go."
• The connector list goes second, and it is the real tier decision. Free connects PostgreSQL, MySQL, ClickHouse, MariaDB and MindsDB. BigQuery, Snowflake, Redshift, Athena, SQL Server / Azure SQL, MongoDB, Databricks, Trino and Dremio are all Team-and-above. If your data is in a warehouse rather than an OLTP database, the Free plan cannot reach it and no amount of compute headroom compensates.
• The machine turns itself off after 15 minutes idle on Free, against 24 hours on Team. Deepnote's long-running-jobs documentation adds a second limit that is not on the pricing page: "Hardware of projects running on the Free plan gets turned off automatically after running continuously for 8 hours", against a "runtime limit of 48 hours" on Team and above. If you run long jobs, that pair of numbers matters more than the RAM figure.
• Three editors, five projects, seven days of history. Viewers are unlimited on every plan, which is genuinely useful — you can share read-only work with as many people as you like without paying. Deepnote's downgrade documentation states the same caps from the other direction: "the workspace will revert to a member limit of 3" and "you will not be able to open any versions older than 7 days".
• Scheduling is not on Free. Deepnote's docs say scheduling is a paid feature and that "Only one scheduled notebook per project is supported"; the downgrade page confirms Free keeps at most one scheduled notebook.
• What Free does not cap: viewers, and Basic machine hours — subject to the $112 ambiguity above. On both of those it is more generous than most free tiers we have looked at.

Deepnote is a notebook. That is the real buying decision.

Deepnote describes itself three different ways on three different surfaces, and the drift tells you something useful about where the product is heading. Its documentation landing page says "Deepnote is the AI workspace for data professionals". Its homepage title tag still says "Deepnote: Collaborative analytics & data science notebook". Its current homepage hero says "The open data workspace built for humans" with "humans" and "agents" alternating.

All three are true, and the notebook one is the one that should drive your decision. Deepnote is where you write Python and SQL, collaboratively, with real compute behind it — scheduled runs, GPUs, model training, a full Jupyter-compatible environment. Deepnote AI sits on top of that as an assistant to the person writing the code. If you have people who write code, that is a strong product and the AI quota on Free is a demo of a feature rather than the whole point of the tool.

If nobody on your side writes Python, the calculation is different, and it is not really a pricing question any more. A notebook with an AI assistant still assumes a notebook user. That is the line where a chat-first tool becomes the relevant comparison — not because it is better, but because it is a different job.

How Intellrise prices the same job

• Free — $0, permanently. One database source plus Google Sheets, 3 dashboards with unlimited charts, unlimited reports and every export format (CSV, Excel, PDF, PPTX, Word). You bring your own AI key, so there is no completion or call quota on the free tier — your allowance is whatever you buy from your provider.
• Pro — $29/mo, or about $24/mo billed annually. Unlimited data sources — the count never enters the price — plus CSV and Excel upload, unlimited dashboards with full customisation, AI Semantic Schema Learning, per-key AI configuration, read-only share links and one-off email delivery.
• Every new account starts on a 14-day Pro trial with no card, and the free tier is the landing state afterwards rather than an expiring demo.
• Bring-your-own-key on every tier — Gemini, OpenAI, Anthropic, DeepSeek, MiniMax or any OpenAI-compatible endpoint. There is no credit meter and no per-question markup; you pay your provider directly for tokens.
• The honest cost comparison at one seat: Intellrise Pro is $29/mo plus tokens; Deepnote Team is $49/mo monthly or $39/mo annual per editor, with $39 of AI credits included. At five seats the gap widens sharply, because Deepnote charges per editor and Intellrise currently has no multi-seat workspace at all — which is a limitation on our side, not a saving.

Where Deepnote is the better buy — and it often is

So here is the concession, and it is large.

If anyone on your team writes Python, Deepnote does things Intellrise cannot do at any price. Scheduled notebook runs, GPU machines, model training and serving, a real Jupyter-compatible environment, Git directory sync, an MCP server, and 40-plus documented integrations including things we have never connected to. Intellrise has no notebook, no scheduling, no GPU and no Python execution. If your work is machine learning or engineering rather than answering business questions, this is not a close call.

Deepnote also beats us on collaboration outright. Unlimited viewers on every tier including Free, real-time multiplayer editing, permission groups on Enterprise, and a multi-seat workspace that actually exists. Intellrise does not have a shared team workspace shipped today, so if more than one person needs the same sources right now, that alone decides it.

And Deepnote's free tier is free forever for students, teachers and researchers with a school email — a genuinely good offer we do not match.

Where the flat-plus-your-own-key shape wins is narrower and worth stating precisely: when the people who need answers do not write code, when the cost of asking a follow-up question should be zero rather than a metered call, and when the deliverable is a dashboard or an exportable report rather than a notebook. Outside that, buy the notebook.

Check it on your own data before you pay for either

Deepnote's Team tier has a 14-day free trial and its Free tier is genuinely usable for a Postgres or MySQL project, so you can evaluate it properly without paying. Do that with your real database rather than a sample, because the connector line is the thing most likely to surprise you.

On our side there is a public demo at https://intellrise.com/demo/ that needs no signup, no card and no AI key — a synthetic sample dataset, the exact SQL each chart runs and the exact numbers it returns — so you can judge the substance before connecting anything. The page is explicit that those queries were hand-written for it rather than generated by the product, so it shows you the output, not the text-to-SQL step. If it looks right, every new account starts on a 14-day Pro trial with no card, and the free tier keeps working afterwards: one database source plus Google Sheets, three dashboards, and unlimited reports and exports on your own AI key.

Two honest caveats on ours. Bring-your-own-key means a setup step: you have to create a provider key before anything runs. And the free tier has no CSV or Excel upload — if your data arrives as files rather than living in a database, our free tier is the wrong tool and Deepnote's is better, though every new account starts with 14 days of Pro, which does upload files.