Datarails is an Excel-connected FP&A, consolidation and reporting platform, now sold as FinanceOS, for mid-market finance teams that want to keep their Excel models while pulling ERP actuals into them and letting AI tools such as Claude query the same governed numbers.
Independently scored, one dimension at a time. Every dimension is sourced, and we say whether a claim is published, reported, or estimated.
Written by the Audit Friendly research team. No vendor edits, no sponsored placement.
What you are really buying: a governed data layer (FinanceOS) under your existing Excel models, with consolidation, permissions, version control, workflows and dashboards on every plan, plus a hosted connector that Datarails says works with Claude, ChatGPT, Copilot, Cursor and other compatible AI tools; Claude's own directory lists it as Anthropic verified. Cash, Month-End Close, Spend and Desk modules sit on that same layer, and since an early October 2026 repackaging (reported by PricingSaaS) the Expert plan bundles all four, with Spend included from Premium up.
The catch: plans cap integrations at one, two or three and named users at 2, 5 or 15, the AI connector only reads, and the pricing page's no-consultant-fees promise sits beside a one-time onboarding fee that Datarails' own 12 October 2026 cost guide discloses without an amount. Datarails' product, pricing and legal pages state no SOC 2 report type (press coverage reports SOC 2 and ISO 27001, and the site links a Trust Center), and modeling depth is whatever your Excel workbooks already do.
Confirm in writing before signing: the onboarding fee and what it covers, the price of each extra user and integration, whether any usage meter applies to the AI connector or agents, the renewal term, notice period and price cap (the Terms of Service leave all three to the Order Form), the SOC 2 report type and the SLA, and how you export mappings and dashboards inside the 45-day post-termination download window.
Jobs, mapped to how finance teams actually buy.
Forget the 40-row feature grid. Pick a finance workflow and see how this tool and its closest peers actually do it, with our verdict on who wins for whom.
Here's what it really costs: published rates where a vendor lists them, aggregated quote data where they don't, plus a path to your own number.
Scored on what an outside agent can actually reach: the documented API, how it authenticates, and whether the vendor ships an MCP server.
Exists: Yes. Datarails runs a hosted MCP server at mcp.datarails.com/mcp, which publishes OAuth protected-resource metadata. Claude's connector directory lists Datarails FinanceOS, made by Datarails, as Anthropic verified and added in August 2026 (checked 10 November 2026). A Claude plugin on GitHub (v3.7.5, released 25 October 2026) bundles the connector with finance skills for Claude Desktop, Cowork and Claude Code. No Datarails server appears in the official MCP registry (checked 10 November 2026).
Scope: Read only. The directory lists 24 tools that list data models and business metrics, fetch rows, run aggregations and distinct-value queries, profile fields, return currency rates and company profile sections, list Filebox contents and return org users. The listing says the connector never creates, modifies or deletes financial data. Raw row reads cap at 500 rows within a 50,000-character response, while grouped calculations run in Datarails' query engine with no row limit, per Datarails' 12 October 2026 article.
Turnkey: Mostly. There is nothing to host: each person signs in with their own Datarails login over OAuth 2.1, and table, row and organization permissions are applied on the server. The connector must first be enabled for the organization. The pricing page lists the FinanceOS AI Connector on all three plans, and Datarails' cost guide says every plan includes it; March 2026 press coverage reported usage-based pricing that the pricing page does not explain.
In-product AI: Datarails sells Strategy, Planning and Reporting agents (available since 21 January 2026), Insights summaries and AI Storyboards for board decks. Datarails says its connector handled more than 41,000 tool calls in the week ending 16 September 2026, with 96 percent completing cleanly; those are the vendor's own production figures, not an independent test.
Who wins for whom: Teams keeping driver logic in their own Excel workbooks that want consolidation and budget submission underneath: Datarails. Excel-first teams that want driver-based and employee-level workforce plans held in a central database: Vena. Spreadsheet teams that want headcount Tables and AI-built scenarios over an OLAP cube: Cube.
Your Excel model, governed underneath. Datarails keeps budgets and forecasts in your own Excel workbooks, adds allocations, currency translation and intercompany eliminations on every plan, and its Planning Agent runs ad-hoc scenarios on the governed data.
Where it loses: No dedicated workforce or driver-based planning module appears on its product pages, and Vendr says it does not natively support workforce planning, so driver logic stays in your spreadsheets.
Source: AF score rationale: Modeling depth 60; Datarails watch-outs.
Watch: How to stop chasing down department budgets one by one using Datarails 1:11 · Sponsored walkthrough · Excel Dictionary on YouTube (Datarails-sponsored)
Excel front end, OLAP cube behind. Vena keeps the full Excel front end over a relational and in-memory OLAP database, with driver-based plans, what-if scenarios and employee-level workforce planning that carries US, Canadian and UK payroll tax logic.
Where it loses: Three-statement links depend on how each template is built, Vena publishes no model size or dimension limits, and a Capterra reviewer reports slowdowns when many users work in large reports.
Source: AF score rationale: Modeling depth 81; Vena watch-outs.
Watch: Discover how Vena helps with financial planning, including a live demo 9:37 · Consultant walkthrough · McCoy TV on YouTube
Scenarios over your spreadsheet model. Cube stores data in an OLAP cube with scenarios, formulas and Tables for headcount rosters and drivers, shows 13-week and rolling forecast views, and its Planner Agent builds new scenarios from plain-language prompts.
Where it loses: Three-statement linking and most driver logic still live in your spreadsheet model, and adding or changing dimensions, including the Time dimension, goes through Cube's support team.
Source: AF score rationale: Modeling depth 70; Cube watch-outs.
Watch: WBSP909: Scale Growth by Understanding Cube FP&A Platform's Capabilities, an Objective Panel Review 1:14:43 · Review · ElevatIQ on YouTube
Who wins for whom: Teams with one to three source systems that want actuals refreshing inside existing Excel reports: Datarails. Business Central or Dynamics 365 F&O shops that want a direct connector with drill-through: Vena. NetSuite teams that want a native saved-search connector with on-demand imports: Cube.
600+ sources, one to three per plan. Datarails lists 600+ sources, with NetSuite, QuickBooks, Sage Intacct, Business Central, Xero, Salesforce and Snowflake marked popular, and refreshes data into existing Excel reports on demand or on a schedule.
Where it loses: Plans include one, two or three integrations, so a company with an ERP, a CRM, an HRIS and a billing system outgrows Professional quickly, and connector direction is not stated on its public pages.
Source: AF score rationale: ERP & data connectivity 71; Datarails watch-outs.
Watch: Datarails: High level demo 12:50 · Official demo · Datarails on YouTube (official)
Direct Business Central and Fabric connectors. Vena documents a direct Business Central connector on API v2.0 with one-click refresh and drill-through, has six generally available Microsoft Fabric connectors including Dynamics 365 F&O, and lists NetSuite, Sage Intacct, QuickBooks, Xero and SAP.
Where it loses: The integrations page does not say which sources use Vena-built connectors versus Fabric pipelines or flat files, and pushing data out through the Export API costs extra on both plans.
Source: AF score rationale: ERP & data connectivity 77; Vena use cases and pricing tiers.
Watch: Northexis - Vena Solutions in 7 Minutes 6:56 · Consultant walkthrough · Northexis on YouTube
Native NetSuite, scheduled imports. Cube's native NetSuite connector reads saved searches over token auth, guides cover Sage Intacct, QuickBooks, Xero, Business Central, Workday, Snowflake and Databricks, and imports run on a schedule or on demand.
Where it loses: Nothing writes back to source systems, and Snowflake, on-premise databases and QuickBooks Desktop can need a Boomi Atom your IT team hosts and keeps online during imports.
Source: AF score rationale: ERP & data connectivity 77; Cube watch-outs.
Watch: Build your monthly cycle once, refresh it forever: Cube product updates | Webinar 43:18 · Official demo · Cube AI on YouTube (official)
Who wins for whom: Teams that want Claude reading governed actuals on any plan, with no write path: Datarails. Vena customers who want Claude reading models and reports while write-back stays in Excel: Vena. Teams that want Claude to draft revised figures into a scenario opened for AI: Cube.
Anthropic-verified, read-only connector. Datarails' hosted MCP server is listed as Anthropic verified in Claude's directory, with 24 tools that read tables, rows, aggregations and KPIs under each user's own permissions, and it sits on every plan.
Where it loses: No tool writes, so Claude cannot update a forecast through the connector, and raw row reads stop at 500 per call, per Datarails' own October 2026 article.
Source: AF score rationale: AI / agent readiness 81; Datarails watch-outs.
Watch: Datarails AI Agents 1:04 · Official demo · Datarails on YouTube (official)
Read-only MCP, write-back in Excel. Vena's hosted MCP server, generally available since 13 May 2026, gives Claude 8 read tools over models, hierarchies, report data and transaction tables within each user's Vena permissions.
Where it loses: Claude cannot write numbers, create scenarios or run workflows through MCP; saving back goes through the Vena Excel add-in, and setup needs an admin.
Source: AF score rationale: AI / agent readiness 80; Vena watch-outs.
Watch: Vena MCP Server: AI Meets Your Governed Data Layer 1:14 · Official short · Vena Solutions on YouTube (official)
Reads, and writes into AI scenarios. Cube's hosted, generally available MCP server documents 30 tools, including 12 build and write tools, so Claude can pull budget versus actual, drill to NetSuite transactions and publish revised figures into a scenario.
Where it loses: Writes need a separate Use MCP (Write) permission and land only in scenarios opened for AI, and AI usage is metered on a credit pool whose pack prices are not public.
Source: AF score rationale: AI / agent readiness 88; Cube watch-outs.
Watch: Cube explained: the agentic FP&A platform for finance teams 1:58 · Official demo · Cube AI on YouTube (official)
AI / agent readiness
Datarails runs a hosted first-party MCP server at mcp.datarails.com, listed as Anthropic verified in Claude's connector directory with 24 tools that the listing calls read-only (checked 10 November 2026). The connector and the Strategy, Planning and Reporting agents are on every plan per the pricing page, and a Claude plugin (v3.7.5, 25 October 2026) adds finance skills. Points come off because no write tools exist and the server is absent from the official MCP registry.
Implementation
Datarails says FP&A implementations average 4 to 6 weeks with no code or IT (FP&A page, 10 November 2026), delivered by its own team for a one-time onboarding fee (cost guide, 12 October 2026). CFO Shortlist (updated August 2026) reports 2 weeks on a fast path to about 12 weeks at full scope. Excel-native models avoid a rebuild; chart-of-accounts mapping still sits with finance.
Support
Support is included in the subscription: standard on Professional, premium on Premium and Expert, with a dedicated customer success manager from implementation onward (Datarails cost guide, 12 October 2026). The homepage shows G2 4.9 and Capterra 4.8 badges, which are the vendor's own display. Vendr reports that premium support tiers can add 10 to 20 percent of the base subscription in some deals, and independent review detail is thin.
ERP & data connectivity
Datarails lists 600+ sources, with NetSuite, QuickBooks, Sage Intacct, Business Central, Xero, Salesforce and Snowflake marked popular on its integrations page, and data refreshes on demand or on a schedule (FP&A page, 10 November 2026). The catch is the cap of one to three integrations per plan. Datarails' public product pages do not state connector direction or per-source refresh intervals, and Workato is listed as an integration sub-processor.
Compliance depth
Datarails' pricing page (checked 10 November 2026) lists version control, approval workflows, a submit button, file boxes and permissions down to entity, department and report on every plan. Its sub-processor list (updated 15 October 2026) names Microsoft Azure for hosting and AI features, with processing in the US, Canada, the UK and the EU (Sweden); it does not say whether a customer picks the region. Its product, pricing and legal pages state no SOC 2 report type (CFOtech reported SOC 2 and ISO 27001 on 12 March 2026, and the site links a Trust Center), and no SOX-specific controls are documented, which holds the score down.
Scalability
Plans cap named users at 2, 5 or 15 (plus 50 viewers on Expert) and integrations at 1, 2 or 3 (pricing page, 10 November 2026). The AI connector caps raw reads at 500 rows within a 50,000-character response but runs grouped calculations with no row limit, per Datarails' 12 October 2026 article. No limits on entities, dimensions or model size are published, and Vendr places mid-market deployments at 15 to 30 users.
Modeling depth
Model logic stays in Excel; Datarails adds consolidation, allocations, currency translation, intercompany eliminations and JV accounting on every plan (pricing page, 10 November 2026). The Planning Agent runs ad-hoc scenarios and the FP&A page lists rolling forecasts. No dedicated workforce or driver-based planning module appears on product pages, and Vendr says Datarails does not natively support workforce planning.
Pricing transparency
Every plan says Request Quote, but the pricing page prints user and integration caps per plan, and Datarails' 12 October 2026 cost guide states typical spend of $35,000 to $80,000 a year for mid-market teams and above $100,000 for enterprise. The same guide discloses a one-time onboarding fee with no amount. Vendr publishes a $33,300 median (updated February 2026) and reports 3 to 5 percent annual escalators.
Datarails' case study with 100 Percent Eyewear says its three-person finance team on NetSuite moved from 3 to 5 basic statements to 25+ detailed reports, refreshed from the Excel add-in with one click. The vendor says use then spread to 30 to 35 people across sales, purchasing and executives.
Datarails' customer stories include ATN Corp cutting multi-entity consolidation from one week to one day and Essence Group unifying finance across different ERPs; both are the vendor's own case studies (customers page, checked 10 November 2026). Multiple charts of accounts, currency translation and intercompany eliminations are on every plan.
Workflows and a Submit button let budget owners send numbers from their own workbooks (pricing page). Datarails' customer stories include United Electric collecting 22 department budgets in one day and Origin Materials running 25 budgets with a lean team, both the vendor's figures.
Datarails' case study with La Fosse says the CFO's 19-person finance team activated the FinanceOS AI Connector in 15 minutes, then had Claude build a 45-page quarterly business review deck from live data. The outcomes are the vendor's own figures; the connector reads data and writes nothing back.
Datarails' case study with ISG says the team left Vena, now uploads raw data and transforms it in the Data Mapper, and halved compensation processing for up to 75 employees on 21 plans. The time saving is the vendor's figure.
Datarails' case study with Tangoe says its 35-person finance team evaluated Adaptive, Jedox and Vena, chose Datarails, and cut month-end reporting from three weeks to three days. The figure is the vendor's.
New Year, New Close: How Real Estate Finance Teams Can Automate Month-End
A 28-minute webinar from REdirect Consulting on automating month-end close for real estate finance teams with Datarails. Relevant to property and construction groups with many entities. Posted 2 February 2026 with under 100 views; it comes from a consultancy, and its commercial relationship with Datarails is not stated in the title.
Datarails AI Agents
Datarails' own one-minute overview of its AI agents, posted 22 February 2026, a month after the Strategy, Planning and Reporting agents became available. Useful for seeing what the in-product agents are meant to do. About 170 views; agent outcomes shown are the vendor's claims.
Datarails Consolidation
Datarails' own two-minute clip on its consolidation capability, posted 10 March 2026, the day Fortune reported the FinanceOS launch. Relevant to multi-entity groups weighing Datarails against a dedicated consolidation tool. About 100 views; it is a vendor feature overview, not a worked example on your own entity structure.
Datarails Live Demo! A financial FP&A platform
A 43-minute live Datarails demo hosted by The Financial Executives Networking Group for its finance-executive audience. The live format suits a buying committee that wants more than a sales clip. Posted 12 September 2025 with about 5,700 views; it predates FinanceOS and the March 2026 AI connector.
Datarails: High level demo
Datarails' own 13-minute high-level demo of the platform. Good for a first look at how the Excel-connected product fits together before a sales call. Posted 15 October 2024 with about 15,000 views; it predates FinanceOS, the AI connector and the 2026 plan structure.
Prepare your PowerPoint presentations swiftly and efficiently.
A 51-second Datarails short on preparing PowerPoint presentations quickly from Datarails data. Relevant to FP&A teams that rebuild monthly board decks by hand. Posted 19 October 2025 with about 1,200 views; it is vendor marketing and too short to show how much deck editing remains.
Excel vs Excel with Datarails
Datarails' own 41-second comparison of working in plain Excel against Excel with Datarails connected. It frames the core pitch for finance teams that do not want to leave spreadsheets. Posted 16 September 2025 with about 4,000 views; it is vendor marketing and predates FinanceOS.
Learn how to create live AI reports in Excel using datarails!
A 52-second Excelbychris short on creating live AI reports in Excel with Datarails. A quick look at the AI layer for teams that want AI inside their existing workbooks. Posted 26 May 2026 with about 29,000 views; output quality in a short clip is untested.
How to create live ERP reports in Excel with Datarails
A 37-second short from the Excelbychris creator channel on building live ERP-connected reports in Excel with Datarails. It suits buyers deciding whether an Excel add-in can replace manual ERP exports. Posted 16 June 2026 with about 30,000 views; a creator demo is no substitute for testing your own ERP connector.
How to automate your FP&A reporting and analysis
A 51-second Excel Dictionary short on automating FP&A reporting and analysis with Datarails. It gives analysts who live in Excel a quick sense of the workflow. Posted 27 June 2025 with about 36,000 views; it predates FinanceOS, so screens may differ from what is sold today.
How to stop chasing down department budgets one by one using Datarails
A 71-second Excel Dictionary short on collecting department budgets through Datarails instead of chasing owners by email. Relevant to FP&A leads who run an annual budget cycle in Excel. Posted 21 September 2026 with about 700 views. A clip this short cannot show approval routing or version control in any depth.
How to automatically roll forward your monthly P&L report using Datarails
A 45-second short from the Excel Dictionary creator channel on rolling a monthly P&L report forward with Datarails instead of rebuilding it each period. Useful for a controller who owns a recurring reporting pack. Posted 13 July 2026 with about 34,000 views. Treat it as promotional: it is a creator walkthrough, not a customer account, and too short to show mapping effort.
Compliance depth
Version control, workflows and permissions on all plans; Azure hosting in US, Canada, UK or EU. SOC 2 reported in press; get the report type from its Trust Center.
Scalability
Plans cap users at 2, 5 or 15 plus 50 viewers and integrations at 1 to 3; AI connector reads cap at 500 rows, grouped calculations are uncapped.
AI / agent readiness
Hosted first-party MCP, Anthropic verified in Claude's directory, 24 read-only tools; connector and AI agents listed on every plan (10 Nov 2026).
Support
Standard support on Professional, premium on Premium and Expert, dedicated CSM from onboarding; vendor displays G2 4.9 and Capterra 4.8.
Implementation
Vendor says 4 to 6 weeks average go-live by its own team for a one-time fee; CFO Shortlist reports 2 weeks fast path to about 12 weeks full scope.
Pricing transparency
Quote only; vendor publishes $35k to $80k a year typical mid-market spend; Vendr median $33,300 a year (page updated February 2026).
Modeling depth
Excel-native models, ad-hoc scenarios via the Planning Agent, rolling forecasts; no dedicated workforce planning module on product pages (10 Nov 2026).
ERP & data connectivity
600+ listed sources incl. NetSuite, Sage Intacct, QuickBooks, Xero, Business Central, Snowflake; refresh on demand or on a schedule.
AF Score
Audit Friendly score for Datarails across eight dimensions, 10 November 2026.
Median contract
Vendr median annual contract value for Datarails, page last updated February 2026, checked 10 November 2026.
Go-live
Average FP&A time to go live that Datarails publishes on its FP&A page, checked 10 November 2026.
Professional
FinanceOS reporting, planning and dashboards, 2 users, 1 integration, Datarails AI Agents, FinanceOS AI Connector, standard support (pricing page, 10 November 2026).
Premium
Everything in Professional plus Spend (cards, bill pay, contracts), 5 users, 2 integrations and premium support. Marked most popular; the feature table marks AI Storyboards as absent only from Professional.
Expert
Everything in Premium plus Cash Management, Month-End Close, Receivables and Desk, 15 users plus 50 viewers, 3 integrations, premium support.
Datarails Cash
Cash position, forecasting and scenario module; an add-on below Expert, bundled in Expert. Datarails says it goes live 2 to 3 weeks after bank connection.
Datarails Month-End Close
Close checklist and sign-off workflow module; an add-on below Expert, bundled in Expert.
Datarails Spend
Requests, purchase orders, contracts, cards, bill pay and reconciliation; bundled from Premium, an add-on to Professional.
Datarails Desk
Finance ticketing with AI-drafted responses and full auditability; PricingSaaS reports it was added in August 2026. Bundled in Expert.
FinanceOS AI Connector
Listed on all three plans and in Datarails' cost guide; must be enabled for the organization. No usage meter is published.
Onboarding
Delivered by Datarails' own team, covering data mapping and builds. Amount not published; Vendr reports $10,000 to $40,000 or more.
Support and CSM
Standard support on Professional, premium on Premium and Expert, dedicated CSM from implementation onward (cost guide, 12 October 2026).
600+ connectors, but your plan includes one to three
Datarails says FinanceOS connects to 600+ systems (FinanceOS page, checked 10 November 2026), yet the pricing page caps Professional at 1 integration, Premium at 2 and Expert at 3. A team with an ERP, a CRM and a payroll system already needs Expert. Price extra integrations before signing; Vendr reports that some custom or legacy connectors carry additional fees or services work.
The AI connector reads; it cannot write
Claude's connector directory lists 24 Datarails tools and says the connector never creates, modifies or deletes financial data (checked 10 November 2026). Forecast updates, mappings and approvals stay in Datarails and Excel. The Claude plugin's Excel agent can drive the Datarails add-in on an open workbook, including publish and submit, so set a policy on who may use it.
No consultant fees, but there is an onboarding fee
The pricing page promises no consultant fees, and Datarails' cost guide (12 October 2026) clarifies that it charges a one-time onboarding fee instead, delivered by its own team and scoped to your setup. The amount is not published. Vendr reports implementation fees of $10,000 to $40,000 or more and says multi-year commitments are the usual lever to reduce or waive them.
Renewal terms live in the Order Form
Datarails' Terms of Service (last updated 15 March 2026) leave the term, fees and payment terms to the Order Form and publish no auto-renewal rule, notice period or renewal price cap. Vendr reports annual escalators of typically 3 to 5 percent. Liability is capped at fees received in the prior 12 months, and customers get 45 days after termination to download their content.
Usage-based pricing is reported, not published
Fortune (10 March 2026) reported that Datarails is shifting to a usage-based pricing model, and CFOtech (12 March 2026) said the finance MCP uses one. The pricing page checked 10 November 2026 still sells plans by users and integrations and names no usage meter. Ask whether AI connector or agent usage is metered now or at renewal.
Security reports sit in the Trust Center, not the product pages
Datarails' product, pricing and legal pages state no SOC 1 or SOC 2 report type and no ISO 27001 certificate, though the site footer links a Trust Center and CFOtech (12 March 2026) reported that Datarails highlights SOC 2, GDPR and ISO 27001. The SLA referenced in the Terms of Service is a PDF, so its uptime commitment and service credits are not on a web page. Request the SOC 2 report, the SLA and the DPA early in procurement.
Model depth is your workbook's depth
Datarails keeps model logic in Excel and adds data connections, consolidation, version control and permissions underneath (cost guide, 12 October 2026). No dedicated workforce or driver-based planning module appears on its product pages, and Vendr says Planful offers workforce planning and advanced consolidation that Datarails does not natively support. Teams without a strong Excel modeler get reports and dashboards, not driver logic.
Agents see 500 rows at a time
Datarails' connector article (12 October 2026) says raw row reads are capped at 500 rows inside a 50,000-character response, while grouped calculations run in its query engine with no row limit. The plugin README says row-by-row extraction pages 500 rows at a time and can take minutes on large tables. Design agent questions around totals, not transaction dumps.
Claude plugin v3.7.5
Datarails released version 3.7.5 of its Finance OS plugin for Claude Desktop, Cowork and Claude Code on GitHub. The current README says the plugin bundles the Datarails connector with skills for financial summaries, variance workbooks, anomaly checks, audit-support packages and an Excel agent that drives the Datarails add-in.
AI connector limits and production figures published
Datarails documented a 500-row cap on raw reads within a 50,000-character response, uncapped grouped calculations in its query engine, and per-user OAuth 2.1 sign-in. It says the connector handled more than 41,000 tool calls in the week ending 16 September 2026 (vendor figure).
Datarails publishes typical spend ranges and an onboarding fee
Datarails' cost guide says most mid-market teams invest $35,000 to $80,000 a year, enterprise plans typically start above $100,000, and implementation is a one-time onboarding fee delivered by its own team. It also says every plan includes the AI agents and the FinanceOS AI Connector.
Plans repackaged: Expert bundles Cash, Close, Receivables and Desk
PricingSaaS reports that the week of 5 October 2026 Datarails added 50 viewers to Expert, moved Spend onto Premium, and made Expert bundle Cash Management, Month-End Close, Receivables and Desk, where it previously included one additional product.
Claude plugin v3.5.0
Datarails released version 3.5.0 of its Finance OS plugin for Claude on GitHub, installable through Cowork, a ZIP upload or Claude Code.
Plans rebranded to FinanceOS; Desk add-on launched
PricingSaaS reports that the week of 24 August 2026 Datarails renamed its plans around FinanceOS, published its first plan comparison table, and added Datarails Desk, a finance ticketing module, alongside Cash, Month-End Close and Spend Control.
Can Datarails replace an existing Excel model?
It does not try to. Datarails' cost guide (12 October 2026) says teams keep their models and formulas in Excel while FinanceOS handles data connections, consolidation, version control and permissions. Driver logic, headcount math and scenario design stay as deep as the workbook you build.
What it replaces is the copy and paste: the Excel add-in refreshes actuals into existing reports, the Data Mapper keeps GL accounts aligned to your reporting structure, and the Planning Agent runs ad-hoc scenarios. If you need a dedicated workforce or driver-based engine, Vendr says Datarails does not natively support workforce planning, so test it in a demo.
Which ERPs and data sources does it connect to, and how often does data refresh?
Datarails lists 600+ sources. Its integrations page marks Oracle NetSuite, QuickBooks (Desktop and Online), Sage Intacct, Microsoft Business Central, Salesforce, Xero, SQL Server, Snowflake, HubSpot, SAP Business One and Acumatica as popular, and also lists Workday, Oracle Fusion Cloud ERP, SAP ECC, BigQuery, Redshift and Power BI (checked 10 November 2026). Custom API connections are available, per the cost guide.
Data refreshes at the click of a button or on an automatic schedule, per the FP&A page; per-source refresh intervals are not published. Each plan includes only one, two or three integrations, so count your systems before choosing a tier. Workato is listed as a sub-processor for integration between customer systems and Datarails.
What do the AI connector and AI agents cost, and which plan includes them?
The pricing page lists Datarails AI Agents and the FinanceOS AI Connector on the entry Professional plan, so all three plans include them, and Datarails' cost guide (12 October 2026) says the same. No separate price is published. PricingSaaS reports that both AI features moved onto all plans in the week of 4 May 2026.
Two caveats. The connector must be enabled for the organization before anyone can use it, per Datarails' 12 October 2026 connector article. And Fortune and CFOtech reported in March 2026 that FinanceOS uses a usage-based pricing model, which the pricing page does not explain, so ask whether any usage meter applies.
Can Datarails consolidate entities that run on different ERPs and charts of accounts?
Yes. The pricing page lists multiple charts of accounts, currency translation, intercompany eliminations, allocations and JV investment accounting (NCI, equity and cost methods) on every plan (checked 10 November 2026). Entities can keep their own ledgers: Datarails' case studies include Essence Group unifying different ERPs and ATN Corp cutting multi-entity consolidation from one week to one day, both the vendor's figures.
The limits are commercial and practical. Each source system counts against the one, two or three integration cap, and Vendr says Planful offers advanced consolidation features that Datarails does not natively support. Ask for a demo on your own entity structure, including minority interests and mid-year acquisitions.
Datarails appears in one of our step-by-step walkthroughs of where AI fits in finance work, with what a person still decides at each step.
Datarails is quote only. Its cost guide (last updated 12 October 2026) says most mid-market teams invest $35,000 to $80,000 a year and enterprise plans typically start above $100,000. Vendr's median Datarails contract is $33,300 a year, with the middle 80 percent of its reported deals between $16,231 and $62,037 (page updated February 2026).
Vendr reports small deployments of 5 to 10 users at $25,000 to $50,000 a year and mid-market deployments of 15 to 30 users at $60,000 to $120,000. Plan choice drives the number: Professional includes 2 users and 1 integration, Premium 5 and 2, Expert 15 plus 50 viewers and 3 (pricing page, checked 10 November 2026). Add a one-time onboarding fee, which Vendr reports at $10,000 to $40,000 or more.
Datarails' Terms of Service (last updated 15 March 2026) set the subscription term in the Order Form and publish no auto-renewal rule, notice period or renewal price cap. Either party may terminate for a material breach left uncured 30 days after written notice, and Datarails' total liability is capped at the fees it received in the 12 months before the claim.
After expiry or termination, Datarails gives customers 45 days to download their content, after which it may delete it. Model logic stays in your Excel workbooks; mappings, dashboards and the governed data layer live in Datarails, so export them inside that window. Vendr reports annual escalators of typically 3 to 5 percent and says buyers negotiate caps on multi-year deals.
Datarails says most FP&A implementations go live in 4 to 6 weeks, depending on data and model complexity, with no code or IT required (FP&A page and cost guide, checked 10 November 2026). Its own team does the work for a one-time onboarding fee covering data mapping and the build of reports, plans and workflows, and a dedicated customer success manager stays on afterwards.
CFO Shortlist (updated August 2026) reports 2 weeks on a fast path for budgeting and about 12 weeks at full scope, faster when one owner can make mapping decisions. Datarails Cash goes live 2 to 3 weeks after the bank connection is established, per its Cash page.
Partly. Datarails runs a hosted MCP server at mcp.datarails.com/mcp, listed in Claude's connector directory as made by Datarails and Anthropic verified, with 24 tools. The listing says the connector is read-only and never creates, modifies or deletes financial data (checked 10 November 2026). Each person signs in with their own Datarails login over OAuth 2.1, and table and row permissions apply on the server.
Claude can list tables and metrics, pull filtered rows 500 at a time, run grouped totals and build Excel or PowerPoint outputs with the Datarails plugin (v3.7.5, 25 October 2026). It cannot change data through the connector. The plugin's Excel agent can drive the Datarails Excel add-in on an open workbook, including refresh, publish and submit. Datarails says the connector also works with ChatGPT, Copilot and Cursor.
Answered only from our own published research on this tool, never from general internet noise. If we cannot answer it well, our research agents will dig in and publish a sourced answer.