Databox review: the 2026 agentic relaunch, Genie AI analyst and MCP server, credit-based pricing from $71/month, user caps, CSV-only export and who should skip it.
On September 9, 2026, Databox CEO Peter Caputa IV published Rethinking Analytics For The Agentic Era, and one line in it sums up the whole post: "Databox is now an agentic analytics platform." The pricing page was rewritten the same day. We found no separate announcement of the price change, so anyone who last priced Databox in 2025 is comparing against a product that no longer exists.
The old model was easy to reason about. In January 2026 the pricing page said "All plans include unlimited users", and every data source past your plan's allowance cost $5.60 a month. The new model removes both. Users are capped at 1, 3 or 10 depending on the tier, data sources come in fixed bundles, and outside the Agency plan the only way to get more of either is to move up a plan. In exchange you get AI: Genie, an AI analyst metered in credits, and an MCP server that lets Claude, ChatGPT or n8n query your Databox metrics directly.
Co-founder Davorin Gabrovec listed what shipped on September 14: "Genie, our AI Analyst, Artifacts, Skills, Routines, MCP Connectors, and the Databox MCP." Agents, the part the word "agentic" most suggests, is still marked "Soon" on every plan.
The pricing page opens on annual billing and hides the monthly figure behind a toggle. Annual is a 20% discount, so the monthly price is a quarter higher than the number you see first. These figures were checked on October 20, 2026, in USD.
| Plan | Billed annually | Billed monthly | Users | Data sources | AI credits / month |
|---|---|---|---|---|---|
| Free | $0 | $0 | 1 | 3 | 50 |
| Analyst | $71/mo | $89/mo | 1 | 5 | 150 |
| Team Core | $199/mo | $249/mo | 3 | 10 | 500 |
| Team Scale | $319/mo | $399/mo | 10 | 30 | 1,000 |
| Agency | from $79/mo | from $99/mo | Unlimited | 20 | 300 |
| Custom | Contact sales | ||||
Databox says it "doesn't charge per seat". True, and beside the point. Analyst is a one-user plan, so the day a second person needs a login the bill goes from $71 to $199 a month on annual billing, or from $89 to $249 monthly. That is 2.8 times the price for one colleague.
Data sources are where the new model bites hardest, because of how they are counted: "Each account or property counts as one, so 3 GA4 properties equal 3 data sources." Three brands, each with GA4, Google Ads, Meta Ads and Search Console, is 12 sources. That is two past Team Core's limit. Under the old pricing those two cost $11.20 a month; now they cost a move to Team Scale, $120 a month more on annual billing.
Agency is the exception to the tier ladder. It is self-serve and grows by account packs, each "1 account, 5 data sources, 50 AI credits". The page is inconsistent with itself here: the pack widget says $20 a month and the FAQ says $25, the Agency card says "Starting at $79 /billed annually" and the FAQ says $99 a month. Read each pair as annual versus monthly.
Add-ons sit on top of any plan. A flexible AI credit top-up is listed at $200 on annual billing, branding and white-labeling at $80, and advanced security management at $40; a fiscal calendar, OKRs and 15-minute sync per source are sold separately again.
Every paid plan has a 14-day trial, and it does not ask for a card. The catch comes on day 14: "If no payment details are added by the end of your 14-day trial, your data will stop syncing and scheduled reports and notifications will be paused." What you built stays. It just stops updating, which on a reporting tool amounts to the same thing as being switched off.
The free plan has shrunk to 1 user, 3 data sources, 1 dashboard, 10 custom metrics, 50 AI credits a month and a daily sync. It also leaves out sharing and notifications, CSV export, API access, the MCP server, chat and email support, and custom data integrations, which is what Databox calls pulling from Google Sheets, Excel, SQL databases or your own API. One dashboard nobody else can see is a demo. A freelancer who wants to send a client a link is on Analyst from the first invoice.
Sync is daily on Free and hourly on every paid plan. Faster is a paid 15-minute add-on per source, and Databox's own help page warns that "Availability varies by data source based on vendor API capabilities and rate limits", so check the connectors you care about before paying for it. The catalogue itself has grown: 130+ native integrations today, up from 80+ in March 2025.
Genie launched on March 25, 2026 (that is the archived copy; the original URL now returns a 404). It answers questions about your connected metrics in plain English, builds dashboards from a prompt and explains why a number moved. During early access it was free with daily limits. According to the help centre, that phase "has ended", every request now spends AI credits, and "Unused credits do not roll over".
The allowances show who Databox thinks the AI is for: 150 credits on Analyst, 500 on Team Core, 1,000 on Team Scale. With a top-up listed at $200, the intended move is to size your tier to your questions rather than buy credits as you go. Use the trial for exactly that. Ask the questions you would ask in a normal week and watch the counter, because it is the only way to learn whether 150 credits is a month of use or a week of it.
The MCP server, released in February 2026, is the more useful half for anyone who already works inside Claude or ChatGPT. Those assistants, and automation tools like n8n, query Databox through its semantic layer, so "how did paid search do last month" resolves to the same metric definition your dashboard uses instead of whatever the model infers from a raw export. It is not on the free plan. On Agency, Genie and MCP add "cross-account reporting", the one AI feature that removes a real chore: one question asked across every client instead of twenty dashboards opened in turn.
Leaving is the weak spot. Databox keeps its own copy of your metric history, but its help centre says datasets can only be exported as CSV files, data "cannot be exported directly from the data sources", and a CSV download link lasts 30 days. CSV export is on Team, Agency and Custom only. On Analyst there is no self-serve export at all. Close the account and the data sits in a 30-day retention period before deletion; anything fuller means asking support for an export before you cancel. Cancellations take effect at the end of the billing cycle, so an annual plan you drop in month two runs, and is paid for, until month twelve.
On security, Databox is SOC 2 certified and GDPR and CCPA compliant, and says plainly "we are not currently ISO 27001 certified". Data lives in AWS's US East region, AES-256 at rest and TLS 1.2 or higher in transit. No EU hosting option is mentioned anywhere. A European team whose procurement asks for EU data residency will not get past that questionnaire.
Support is chat and email on paid plans only, Monday to Friday in US Eastern business hours, and there is no formal SLA. The first reply is meant to come within minutes. Resolution is another matter: Databox's support page puts the "median resolution time" at "around 5 business days". That gap shows up in what customers write. Databox holds a TrustScore of 3.2 out of 5 on Trustpilot from 26 reviews; one from July 2026 reads "We have been trying to configure it for over a year, but it breaks constantly", and an April 2026 reviewer says the company "auto-renewed my account after 6 months".
Buy it if you are one marketer, or a team of two or three, pulling from five to ten sources, and you will actually put questions to Genie or wire your metrics into Claude through MCP. Analyst at $71 a month (annual) or Team Core at $199 is fair money for that, and a trial without a card makes it cheap to find out.
Skip it if you are four or more people with many GA4 properties and ad accounts. The user caps and the per-property count push you to Team Scale, $319 a month on annual billing or $399 monthly, before anyone has asked Genie a single question. If all you need is dashboards, price the same setup on DashThis first. Agencies should put their client list through AgencyAnalytics, which charges per client with everything included, before buying Databox account packs. And if your contracts require EU hosting or ISO 27001, Databox cannot pass that check today.
The relaunch made Databox a better product for small teams who want AI on top of their KPIs and a worse deal for everyone who chose it for unlimited users and cheap extra sources. Know which group you are in before the trial ends.
Visit the official website to see the latest plans, pricing, and special offers.
Get Started - Databox* Affiliate link - we may earn a commission at no extra cost to you.