# Layerz

> Layerz is the structured financial modeling layer for AI agents and finance teams. It separates model **structure** (logic, the dependency graph of variables, timelines) from **data** (values, scenarios), so financial models can be versioned, audited, reused, and exported to clean Excel without drift. Claude and other MCP clients drive Layerz from the outside, keeping full context end to end. Layerz is not an Excel generator; Excel is a compatibility output format.

## What Layerz Does

Layerz lets you create structured financial models where every assumption is explicit, every formula is transparent, and every output is traceable. Models are built from typed items (assumptions, formulas, balances, callups, sections) organized in a section-first hierarchy. The compute engine evaluates all items in dependency order, and the export pipeline generates self-contained Excel workbooks with live formulas, editable assumption cells, and proper financial formatting.

Unlike spreadsheets, Layerz models never break when you change structure. Add a timeline, insert an item, modify a formula, and the engine re-resolves all dependencies automatically. The structure is versioned and reusable; the data stays separate.

## For AI Agents: the MCP server

The primary way agents use Layerz is the **remote MCP server**:

- Endpoint: `https://layerz.cc/mcp` (transport: streamable-http)
- Auth: OAuth sign-in flow, or API key
- Exposes 29 tools and 5 resources (FINANCE.md per model, custom instructions, modeling guide, formula and format specs): build, read, patch, compute, validate, diff, version, import, and export models programmatically.
- Minimal client config:

```json
{ "mcpServers": { "layerz": { "url": "https://layerz.cc/mcp" } } }
```

Every model carries an open-standard `FINANCE.md` (its objective + conventions: currency, units, sign, glossary, and the structure of the P&L / Cash Flow / Balance Sheet). Agents should read it first and keep it in sync with the live model.

A REST API and a CLI are also available for non-MCP workflows (see below).

## Use Cases

Layerz is used to build any financial projection model programmatically:

- **SaaS Revenue Model**: MRR, churn rate, expansion revenue, unit economics, LTV/CAC
- **Startup Runway**: Burn rate, funding rounds, hiring plan, cash runway milestones
- **LBO Analysis**: Debt tranches, leverage ratios, interest schedules, equity returns, IRR
- **Manufacturing P&L**: Raw materials, labor costs, production capacity, gross margins, capex
- **Consulting Agency**: Utilization rates, day rates, headcount planning, project pipeline
- **Infrastructure Project**: Capex schedule, depreciation, operating cash flows, IRR/NPV
- **Fintech Lending**: Loan book growth, default rates, net interest margin, provisions
- **Real Estate Development**: Construction costs, rental income, vacancy rates, cap rates
- **E-commerce**: GMV, conversion rates, average order value, fulfillment costs, contribution margin

## Why AI Agents Use Layerz

- **Deterministic & Auditable**: Every formula is transparent. Results are reproducible. No hidden cells.
- **Structured, Not Flat**: Typed items with explicit dependencies. No cell-reference guessing.
- **Persistent Context**: The model, its `FINANCE.md`, and its history persist across sessions. No re-explaining.
- **Compute on Demand**: One call evaluates the entire model in dependency order.
- **Open Excel Output**: Export audit-ready .xlsx with live formulas. No vendor lock-in. Export is never paywalled.
- **Version-Controlled Logic**: Model structure behaves like code. Version it, diff it, branch it.
- **Multi-Timeline**: Monthly, yearly, quarterly, constant periods with automatic cross-timeline projection.

## Key Concepts

- **MDK (Model Definition Kit)**: The pivot format: structured JSON describing a model's logic without data.
- **Model**: Named container with timelines and ordered items. Pre-loaded with default sections (P&L, Cash Flow, Balance Sheet).
- **Timelines**: Four grains: `constant` (single value), `monthly`, `yearly`, `quarterly`. Each has start/end dates.
- **Items**: Roles: `assumption` (input), `formula` (computed, or a no-expression aggregator that sums its children), `callup` (typed mirror of one source, e.g. a balance flow), `balance` (cumulative stock with opening balance + flows), `section` (structural grouping, nestable), `dashboard` (holds `chart` / `kpi` / `mini_table`). Note: the pre-MDK-0.2.0 `layer` role no longer exists.
- **Lists**: Parametric dimensions (e.g. departments, products, debt tranches) via `liste_ref`: one item expands to one instance per element, with cross-list bridging via `mapped_to`.
- **Compute**: Evaluates all items via topological sort; returns values indexed by UID.

## Authentication

- MCP: OAuth sign-in, or API key.
- REST API key format: `lz_<40 hex>`. Pass as `Authorization: Bearer <key>`. Generate keys at layerz.cc → Avatar → Layerz for Agents.
- Rate limit: applied per plan/IP to protect infrastructure (not to meter usage).

## Formula Functions

Scalar: `IF`, `AND`, `OR`, `NOT`, `MAX`, `MIN`, `ABS`, `POW`, `EDATE`, `IN_PERIOD`, `NPV`, `IRR`, `PMT`, `IPMT`, `PPMT`.
Aggregates over a UID or list: `SUM`, `AVG`, `COUNTA`, `SUMIF`, `COUNTIF`, `AVGIF`.
Magic identifier: `CURRENT_PERIOD` (current period end as a date serial) for time-gated patterns (hire ramps, contract activation, depreciation windows).
Lag syntax: `uid_Y-1` (previous year), `uid_M-1` (previous month), `uid_Q-1` (previous quarter); variable lag `_M-$delay`.
List subscript: `` Item[`key`] `` slices a list-dimensioned item by element or by a `mapped_to` key (auto-aggregates).

## Research & Data

Sourced statistics pages. Every figure carries its primary source, sample size and
publication year; figures that could not be verified are listed and explained rather
than dropped silently. Free to cite with attribution.

- [The AI-native fractional CFO, the 2026 Fractional CFO Study report](https://layerz.cc/fcfo-report): Layerz's own survey of 155 fractional CFOs from 35 countries (October 2026) on how they use AI on client work: adoption, controls, the monthly close as a system, day rates by region. Free PDF on request.

- [SaaS Spending Benchmarks: What Private SaaS Companies Spend, by Department and Stage (2026)](https://layerz.cc/resources/saas-spending-benchmarks): What private SaaS companies spend on sales, marketing, R&D and G&A, plus growth, retention and CAC payback by stage. Each figure with its denominator.
- [AI Financial Model Risk and Governance: Statistics (2026)](https://layerz.cc/resources/ai-financial-model-risk-statistics): Verified statistics on AI financial model risk and governance: accuracy benchmarks, shadow AI in finance, audit gaps, and EU AI Act rules.
- [AI in Finance: Adoption and Reality Statistics (2026)](https://layerz.cc/resources/ai-in-finance-statistics): How many finance teams use AI, where agents are heading, and why most enterprise AI efforts still fall flat. The 2026 numbers, with primary sources.
- [E-Invoicing Mandate Statistics (2026)](https://layerz.cc/resources/e-invoicing-mandate-statistics): Dated figures on Europe's e-invoicing mandates: France live 1 Sept 2026, platform registration counts, formats and lifecycle statuses. Sourced.
- [Financial Forecast Accuracy and Budget Cycle Statistics (2026)](https://layerz.cc/resources/forecast-accuracy-and-budget-cycle-statistics): How accurate are corporate forecasts, and how long does the annual budget really take? The data, with primary sources, sample sizes, and years.
- [Fractional CFO Statistics: Demand, Pay, and the Engagement Model (2026)](https://layerz.cc/resources/fractional-cfo-statistics): Fractional and part-time CFO statistics for 2026: demand, pay, clients per CFO, client size, hours, remote work and AI use, each figure with its primary source.
- [Fractional CFOs in SEC Filings: Statistics (2026)](https://layerz.cc/resources/fractional-cfo-sec-filings-statistics): Fractional CFOs in US SEC filings, 2015 to 2026: how the vocabulary changed, how many companies disclose one, and what they pay.
- [LLM Reliability: Hallucination, Consistency, and Reasoning Statistics (2026)](https://layerz.cc/resources/llm-reliability-statistics): How often do LLMs hallucinate, how consistent are they run to run, and where does reasoning break down? Verified LLM reliability data, a source per figure.
- [Spreadsheet Errors and Risks: Statistics and Notable Failures (2026)](https://layerz.cc/resources/spreadsheet-errors-and-risks): How often do spreadsheets contain errors, how much does finance still rely on them, and what have spreadsheet mistakes cost? The data, with primary sources.
- [The Finance Engineer Role: Job Posting Statistics (2026)](https://layerz.cc/resources/finance-engineer-statistics): 418 job boards and 28,087 postings scanned: 27 verified finance engineering requisitions. Titles, departments, tools, pay, and what changed since July.
- [What a Finance Function Costs: Benchmark Statistics (2026)](https://layerz.cc/resources/finance-function-cost-statistics): Finance cost as a share of revenue, finance FTEs per $1bn, and where finance time goes. Benchmark data with primary sources and sample sizes.
- [Where FP&A Time Actually Goes: Productivity Statistics (2026)](https://layerz.cc/resources/fpa-productivity-statistics): How much of FP&A time reaches analysis, and what stops the rest getting there. Verified time-allocation data for financial planning, a source per figure.

## Guides: AI and Financial Modeling

How to build financial models with an LLM and keep them correct. This is the core of what
Layerz writes about: where the model does well, where it breaks, and what a structure layer
changes. Each guide names the failure mode before the fix.

- [ChatGPT Plugin for Financial Models: Layerz in ChatGPT and Codex](https://layerz.cc/blog/chatgpt-plugin-for-financial-models): Layerz is an official ChatGPT plugin for financial models, also in Codex: what it does, how to add it, and what stays under your control.
- [Codex for Finance: How to Build and Edit Financial Models (2026)](https://layerz.cc/blog/codex-for-finance): How finance teams use OpenAI Codex to build and edit financial models: an AGENTS.md for finance, a FINANCE.md per model, a model layer over MCP, Excel out.
- [The CFO Harness: What Sits Between an AI Model and a Number You Sign](https://layerz.cc/blog/cfo-harness): A CFO harness is the layer around an AI model that makes its finance work signable: structured model, conventions, permissions, history and review.
- [Sanity-Check a SaaS Financial Model Against Benchmarks: Five Ratios and How to Read Them](https://layerz.cc/blog/sanity-check-saas-model-against-benchmarks): Five ratios to compute from your own SaaS model, the benchmark ranges to hold them against, and the four mistakes that make the comparison meaningless.
- [Does the AI Do the Math? 10 Questions to Ask Before You Buy an AI Finance Tool](https://layerz.cc/blog/questions-to-ask-ai-finance-tool): Every finance tool now has AI in it. Ten questions that separate the ones you can defend in front of an auditor, and what a good answer sounds like.
- [Semantic Layer vs Model Layer: What a Number Means vs What It Will Become](https://layerz.cc/blog/semantic-layer-vs-model-layer): Finance AI has converged on the semantic layer: metric definitions, lineage, permissions. It answers what a number means, not what it will become.
- [Why Not Just Build It Yourself With AI?](https://layerz.cc/blog/build-it-yourself-with-ai-vs-model-layer): Claude Code can build the finance tool you were about to buy. The honest version of when DIY wins, and the three questions that actually decide it.
- [How to Stop AI From Hardcoding Values in a Financial Model](https://layerz.cc/blog/stop-ai-hardcoding-financial-model): Hardcoded values are the most common way an AI-built model fails, and the hardest to see. Seven checks that catch them, and why prompting harder does not work.
- [E-Invoicing Just Made Your Actuals Machine-Readable. Your Forecast Should Notice.](https://layerz.cc/blog/e-invoicing-forecasting-transaction-level): France's e-invoicing mandate is live. It standardises the invoice, not your access to it, and that decides whether you can forecast from primary data.
- [AI Financial Model Generator: What These Tools Actually Produce](https://layerz.cc/blog/ai-financial-model-generator): Four kinds of AI financial model generator, what the 2026 benchmarks say they get wrong, and the three questions that tell you if the output is usable.
- [Four Ways to Model Inventory in a Financial Model (and What Switching Costs You)](https://layerz.cc/blog/how-to-model-inventory-financial-model): Percent of COGS, days, unit build, ageing from flows. The four methods agree in year 1, land 865k apart by year 5, and the P&L never shows it.
- [Persistent Memory for a Financial Model: The Six Things That Must Survive a New Session](https://layerz.cc/blog/persistent-memory-financial-model-sessions): An AI agent forgets your model between sessions. Six things have to persist for the next session to start where the last one stopped, plus how to test each.
- [FINANCE.md: An Open Standard for Financial Modeling Conventions](https://layerz.cc/blog/finance-md-standard): Code has CLAUDE.md. Design has tokens. Finance still runs on institutional memory. FINANCE.md is one versioned file telling any agent how to read your numbers.
- [Can You Trust an AI-Generated Financial Model? What 63 Expert Modellers Say](https://layerz.cc/blog/can-you-trust-an-ai-generated-financial-model): 63 of the world's leading financial modellers were asked if they'd trust an AI model without review. Zero said yes. What that means for how you build.
- [Claude for Financial Modeling: Where the LLM Ends and the Model Layer Starts](https://layerz.cc/blog/claude-for-financial-modeling): Claude is a superb reasoning engine but a poor calculator. Here's the boundary between what the LLM should do and what a model layer should hold.
- [Token Efficiency for Finance: 8 Habits That Cut Your AI Cost](https://layerz.cc/blog/token-efficiency-finance-ai): The biggest AI cost in finance isn't the answer, it's everything the model re-reads first. Eight habits to cut token cost, and the one that beats them all.
- [Claude Code for Finance: How to Build and Edit Financial Models (2026)](https://layerz.cc/blog/claude-code-for-finance): How finance teams use Claude Code to build and edit financial models: a four-step setup, a CLAUDE.md for finance, and the model layer that holds.
- [How to Build an Auditable Financial Model with AI](https://layerz.cc/blog/build-an-auditable-financial-model): An auditable model answers 'where does this number come from?' in seconds. Here's how to build one with AI, with full dependency traceability both ways.
- [How to Build a Financial Model with AI That Doesn't Drift](https://layerz.cc/blog/financial-model-that-doesnt-drift): AI does 80% of the modeling work, then drifts on the 20% that matters. Here's why models drift under AI, and how to keep the logic intact through edits.
- [How to Save Tokens When Building Financial Models with AI](https://layerz.cc/blog/save-tokens-building-financial-models-with-ai): Claude re-reads your whole model on every edit, which is why tokens vanish in ten prompts. Here's how to build financial models with AI without burning budget.
- [The Right Way to Generate a Financial Model with AI](https://layerz.cc/blog/the-right-way-to-generate-a-financial-model-with-ai): AI generates a financial model in seconds. Most outputs drift, forget, and break. Here's the difference between generating a model and keeping one.
- [Why Finance Agents Need a Model Layer, Not Just Spreadsheet Access](https://layerz.cc/blog/why-finance-agents-need-a-model-layer): Connecting an AI agent to a spreadsheet gives it cells. A model layer gives structure, types, dependencies, and persistence. Here's why the difference matters.

## Guides: Working With Your Accounting and Payment Stack

Your ledger holds the data and hands you a posting batch, not a model. These guides cover the
workflow from source system to a forecast that survives next month. No connector required in
any of them: the export is the interface.

- [Month-End Close Across Multiple Clients: An AI Playbook for Fractional CFOs](https://layerz.cc/blog/fractional-cfo-month-end-close-ai): A month-end close playbook for fractional CFOs with several clients: staggered calendar, acceptance checks, logged adjustments, and where AI fits.
- [The Fractional CFO Tech Stack, Layer by Layer (2026)](https://layerz.cc/blog/fractional-cfo-tech-stack): A fractional CFO tech stack built by layer: ledger, bank, close, model, board pack, AI and MCP. What you carry between clients, and what stays with each client.
- [QuickBooks AI Inventory Forecasting: Exports, Gaps and the Math](https://layerz.cc/blog/quickbooks-inventory-forecasting-ai): QuickBooks knows what you sold, not what you could have sold. Which exports to give Claude, what they hide, and the reorder math that turns them into cash.
- [How to Analyze Your QuickBooks Data with AI](https://layerz.cc/blog/analyze-quickbooks-data-with-ai): QuickBooks holds the history. Getting real analysis out of it with Claude takes the right exports, a fixed account mapping, and a home for the answer.
- [How to Use AI with DATEV for Financial Analysis](https://layerz.cc/blog/how-to-use-ai-with-datev): DATEV holds your books but hands you a posting batch, not a model. Here's how to get real analysis out of it with Claude, and why you need no connector.
- [How to Use AI with QuickBooks for Forecasting](https://layerz.cc/blog/how-to-use-ai-with-quickbooks): Plenty of AI runs on QuickBooks, and almost all of it automates bookkeeping. Here's how to use Claude for the job none of it does: a financial forecast.
- [How to Use AI with Xero for Financial Planning](https://layerz.cc/blog/how-to-use-ai-with-xero): Xero gives you clean data and hundreds of apps, and still leaves you rebuilding your forecast in a spreadsheet. Here's how to use Claude on your numbers.
- [How to Use AI with Pennylane for Financial Analysis](https://layerz.cc/blog/how-to-use-ai-with-pennylane): Hand your Pennylane ledger to Claude and get real financial analysis. Here's the workflow that works, where it breaks, and how to make it last next month.
- [How to Use AI with Qonto for Cash and Runway](https://layerz.cc/blog/how-to-use-ai-with-qonto): Hand your Qonto transactions to Claude and get real analysis. Here's the workflow that works, where it breaks, and how to make the output survive next month.
- [How to Use AI with Stripe for SaaS Revenue Modelling](https://layerz.cc/blog/how-to-use-ai-with-stripe): Hand your Stripe data to Claude and get real revenue analysis. Here's the workflow that works, where it breaks, and how to make the output last.
- [How to Use AI for Your Monthly Financial Close (Without Losing the Audit Trail)](https://layerz.cc/blog/how-to-use-ai-for-monthly-financial-close): A practical, stage-by-stage guide to using AI in the monthly financial close: where it saves time, where it creates risk, and where a human has to sign off.
- [Budget vs. Actuals: Why Finance Teams Lose Hours Every Month to a Problem That Shouldn't Exist](https://layerz.cc/blog/budget-vs-actuals-why-finance-teams-lose-hours): Budget vs. actuals shouldn't take two days every month. Here's the structural root cause, and how finance teams eliminate the monthly reporting grind.

## Guides: Governance, Risk and Data Security

What actually happens to financial data sent to an AI provider, what client contracts forbid,
and what defensible practice looks like. Written for people who have to answer to a board, a
client, or a compliance function.

- [How Fractional CFO Firms Handle Client Confidentiality With AI](https://layerz.cc/blog/fractional-cfo-client-confidentiality-ai): How fractional CFO firms handle client confidentiality with AI: a one-page AI policy per client, an engagement letter clause, and set-ups for strict clients.
- [The Verification Tax: Why Checking an AI Financial Model Costs More Than Building One](https://layerz.cc/blog/ai-financial-model-verification): Close to 13 hours a week go into validating AI outputs, and 71% of finance leaders would reject a 99% accurate tool that cannot explain itself.
- [Your Client's Contract Says No AI. What You Can Still Do.](https://layerz.cc/blog/client-contract-prohibits-ai-financial-data): Client contracts increasingly ban AI on confidential financials. What those clauses forbid, what providers do with your data, and four workflows that comply.
- [AI and Confidential Financial Data: What Actually Happens to Your Numbers](https://layerz.cc/blog/ai-confidential-financial-data-security): Using AI on confidential financial data? Here is what actually happens to your numbers, and why the tool architecture determines your real exposure.
- [Financial Model Audit Trails: What a PE Board Actually Asks For](https://layerz.cc/blog/financial-model-audit-trail-pe-board): A financial model audit trail must show where each number comes from, what changed, who changed it, and why. Here is what passes a PE board review.
- [Shadow IT in Finance: The Hidden Risk of Using AI for Financial Modeling Without a Framework](https://layerz.cc/blog/shadow-it-ai-financial-modeling-risk): Shadow IT in finance is everywhere. Here are the 4 variables to map before using AI on client data, and what real AI governance actually looks like.

## Guides: The Finance Function and the Role

- [Claude Code for Fractional CFOs: One Folder per Client (2026)](https://layerz.cc/blog/claude-code-for-fractional-cfos): How fractional CFOs run several clients in Claude Code: one folder per client, a CLAUDE.md for your method, skills for the close, a model per client over MCP.
- [Codex for Fractional CFOs: One Agent, Several Clients (2026)](https://layerz.cc/blog/codex-for-fractional-cfos): How fractional CFOs run several clients with OpenAI Codex: one folder and one AGENTS.md per client, skills for the close, a bounded model connection each.
- [Fractional CFO vs Part-Time CFO: What Eleven Years of SEC Filings Show](https://layerz.cc/blog/fractional-cfo-vs-part-time-cfo): Fractional CFO or part-time CFO? In US SEC filings, "fractional" went from 5% to 59% of external-CFO mentions since 2015. What changed, and what it pays.
- [ChatGPT for Fractional CFOs: What It Does Well, and What to Add](https://layerz.cc/blog/chatgpt-for-fractional-cfos): How fractional CFOs use ChatGPT across clients: Projects, custom GPTs, data analysis, plugins and MCP, and how to get your practice recommended by ChatGPT.
- [Claude for Fractional CFOs: One Setup, Several Clients (2026)](https://layerz.cc/blog/claude-for-fractional-cfos): How a fractional CFO sets up Claude across several clients: one Project per client, MCP connectors, a FINANCE.md per model, and what to keep out.
- [The AI Fractional CFO: What to Automate, What to Sign (2026 Guide)](https://layerz.cc/blog/ai-for-fractional-cfos): How an AI fractional CFO practice runs in 2026: what fractional CFOs hand to AI, what they keep under their signature, and what AI changes about the job.
- [GitHub for Finance Teams: A Setup Guide for CFOs (2026)](https://layerz.cc/blog/github-for-finance-teams): How to set up GitHub for a finance team: two repos, what goes where, the review loop, the paid-plan detail most guides skip, and the Excel file git cannot read.
- [GitHub for Fractional CFOs: One Method Repo, One Repo per Client](https://layerz.cc/blog/github-for-fractional-cfos): How a fractional CFO sets up GitHub across clients: one private method repo, one repo per client owned by the client, review when alone, and a clean exit.
- [Financial Projections Before Revenue: How to Set Assumptions When There Is No Percentage to Take](https://layerz.cc/blog/financial-projections-before-revenue): Benchmarks say what share of revenue to spend on marketing. At zero revenue that gives nothing. How to build defensible assumptions for a pre-revenue model.
- [Taking Over a Client's Financial Model: A Fractional CFO's First-Week Checklist](https://layerz.cc/blog/fractional-cfo-client-model-takeover): A new client hands you a budget you did not build and the board meets next week. Five steps: map it, find the hardcodes, tie it to actuals, decide what to keep.
- [Fractional CFO Financial Model Template: How to Reuse One Across Every Client](https://layerz.cc/blog/fractional-cfo-financial-model-template): Four clients, four copies of the same 13-week cash template, each fixed differently. What to standardise, what to keep per client, and when copying is fine.
- [Budget 2027: Govern the Commitment Before You Sign It](https://layerz.cc/blog/budget-2027-governing-commitments): The 2027 budget locks in multi-year commitments that used to be variable costs. Five practices that matter, and the one thing none of them work without.
- [The Finance Engineer Job Description Is a Spec for Infrastructure Nobody Has Built](https://layerz.cc/blog/finance-engineer-job-description): We read every finance engineer posting we could verify. They ask one person to own version control, an API surface and audit trails for a financial model.
- [Finance Software Went Closed. Developer Tools Went Open. AI Is Ending the Divide.](https://layerz.cc/blog/finance-software-open-vs-closed): Planning software never had a free tier or an open format. Developer tools had both for years. Here's why AI is forcing finance to converge.
- [Where CFOs Should Actually Start With AI (a Function-by-Function Map)](https://layerz.cc/blog/where-cfos-start-with-ai): A practical map of where AI pays off across the finance function, where it stays supervised, and the one starting point that compounds every month.
- [Will AI Replace the CFO? The Two-Layer Answer](https://layerz.cc/blog/will-ai-replace-cfo): Anthropic puts business and finance at 94.3% AI coverage. But the CFO role isn't disappearing, it's splitting into two layers. Here's the framework.
- [Understanding an Inherited Financial Model Fast: A Framework for M&A Analysts](https://layerz.cc/blog/understanding-inherited-financial-model-fast): Inherited a 200-tab model with a 2-day deadline? This framework helps M&A analysts map structure, flag red flags, and write a defensible review, fast.
- [How to Stop Rebuilding Your M&A Model From Scratch Every Deal](https://layerz.cc/blog/stop-rebuilding-ma-model-from-scratch): Every M&A deal rebuilt from scratch costs your PE-backed finance team days. Learn why separating model structure from data is the fix, and how to start.

## Integrations

How Layerz fits alongside common finance and data tools. Import paths, what maps to what,
and what the workflow looks like end to end.

- [Connect Airtable to Layerz](https://layerz.cc/integrations/airtable): Connect Airtable to Layerz: pull rows from any base into a structured financial model, map columns once, sync on demand. Works with Claude via MCP.
- [Connect Grok Bot to Layerz](https://layerz.cc/integrations/grok-bot): Connect Grok Bot to Layerz over MCP: turn a bank export into a structured financial model, update it from the chat, export Excel with live formulas.
- [Connect Metabase to Layerz](https://layerz.cc/integrations/metabase): Connect Metabase to Layerz: turn a saved question into model actuals, map columns once, sync on demand. Works with Claude via MCP.
- [Connect Pennylane to Layerz](https://layerz.cc/integrations/pennylane): Connect Pennylane to Layerz: pull your ledger into a structured model, branch scenarios and export audit-ready Excel. Works with Claude via MCP.
- [Connect Qonto to Layerz](https://layerz.cc/integrations/qonto): Connect Qonto to Layerz: pull categorised transactions into a structured model, track cash against your plan, export clean Excel. Works with Claude via MCP.
- [Connect Stripe to Layerz](https://layerz.cc/integrations/stripe): Connect Stripe to Layerz: pull real MRR, invoices and payouts into a structured SaaS model, reforecast and export clean Excel. Works with Claude via MCP.
- [Import your DATEV export into Layerz](https://layerz.cc/integrations/datev): Import your DATEV export (SKR03 / SKR04) into Layerz: turn your German ledger into a structured financial model, reforecast and export clean Excel.
- [Import your FEC into Layerz](https://layerz.cc/integrations/fec): Import your FEC into Layerz: turn a full-year French statutory ledger into a structured financial model, branch scenarios and export audit-ready Excel.

## Plans and Facts

- Plans: Individual: free, unlimited models, one editor. Team: €49 excl. tax per seat per month (same amount in USD and GBP), from the second editor.
- History: Individual: 24-hour history and one pinned version. Team: 30-day history, unlimited pinned versions.
- AI: No built-in AI: your own Claude, or any MCP client, drives the model through 29 MCP tools.
- AI clients: Claude, Claude Cowork, Claude Code, ChatGPT, OpenAI Codex, Grok, Grok Bot, Mistral Le Chat, OpenCode, all over OAuth (setup: https://layerz.cc/for-agents)
- Export: .xlsx with live formulas, JSON and raw export on every plan, never paywalled.
- Import: Excel (.xlsx, .xls) and CSV import, with a fidelity score for the formulas kept live.
- Workspaces: Team workspaces, each with shared conventions for AI agents and a template library.
- Conventions: Every model carries a FINANCE.md: currency, units, sign convention and structure, read by the agent.

## Comparisons

Written to say plainly when the other tool is the right answer. A comparison that never
concedes anything is not useful to a reader or to a model summarising the space.

- [Layerz vs Claude](https://layerz.cc/vs/claude): Reviewed 2026-10-19. Claude alone drafts a model in a chat or an .xlsx file; Layerz is not an alternative but a structured spreadsheet Claude drives through MCP, keeping variables, history and conventions between sessions. Free for one editor.
- [Layerz vs Claude in Excel](https://layerz.cc/vs/claude-in-excel): Reviewed 2026-10-19. Claude in Excel: Anthropic's add-in that edits the open workbook, included in paid Claude plans. Layerz: a structured spreadsheet the same Claude drives through MCP, with history and conventions kept between sessions, free for one editor. Many users do both.
- [Layerz vs ChatGPT for Excel](https://layerz.cc/vs/chatgpt-for-excel): Reviewed 2026-10-19. ChatGPT for Excel: OpenAI's add-in that edits the open workbook, on all ChatGPT plans since 5 May 2026. Layerz: a structured spreadsheet an AI agent drives through MCP, with history and conventions kept between sessions, free for one editor.
- [Layerz vs Copilot in Excel](https://layerz.cc/vs/copilot-excel): Reviewed 2026-10-19. Copilot in Excel: Microsoft's AI in Excel, with finance skills and market data connectors on a Microsoft 365 Copilot licence. Layerz: a structured spreadsheet your own Claude drives through MCP, exported to .xlsx, free for one editor.
- [Layerz vs Shortcut](https://layerz.cc/vs/shortcut-ai): Reviewed 2026-10-19. Shortcut: an AI agent for Excel sold on credits (Pro $100 a month billed annually). Layerz: a structured spreadsheet driven by your own Claude through MCP, free for one editor. A Shortcut workbook can be imported into Layerz.
- [Layerz vs Excel](https://layerz.cc/vs/excel): Reviewed 2026-10-19. Excel: the universal spreadsheet, with Copilot and AI add-ins in the workbook. Layerz: a structured spreadsheet an AI agent edits without breaking it, with model-level history, exported to .xlsx with live formulas on every plan.
- [Layerz vs Pigment](https://layerz.cc/vs/pigment): Reviewed 2026-10-19. Pigment: enterprise planning platform with built-in agents and an MCP server, no published price. Layerz: structured spreadsheet driven by your own Claude, free for one editor. Choose Pigment for company-wide planning, Layerz for a model one person builds and hands over.
- [Layerz vs Anaplan](https://layerz.cc/vs/anaplan): Reviewed 2026-10-19. Anaplan: enterprise connected-planning platform with built-in agents (CoModeler, Agent Studio), no published price, no official MCP server found. Layerz: self-serve structured spreadsheet driven by your own Claude, free for one editor.

All comparisons, with the archetype map: https://layerz.cc/vs

## Product Pages

- [FINANCE.md, the open standard](https://layerz.cc/finance-md): Specification and rationale for the conventions file every Layerz model carries.
- [For AI Agents](https://layerz.cc/for-agents): MCP setup, auth, tool surface, and the agent integration guide.
- [Financial Model API](https://layerz.cc/financial-model-api): REST API reference and use cases.
- [For Fractional CFOs](https://layerz.cc/fractional-cfos) · [For Finance Teams](https://layerz.cc/finance-teams) · [For Advisors](https://layerz.cc/advisors) · [Project Finance](https://layerz.cc/project-finance)
- [Trust and Security](https://layerz.cc/trust): Data handling, subprocessors, and what leaves your account.
- [Pricing](https://layerz.cc/pricing): Individual: free, unlimited models, one editor. Team: €49 excl. tax per seat per month (same amount in USD and GBP), from the second editor. Excel export is never paywalled on any plan.
- [Blog index](https://layerz.cc/blog) · [Research index](https://layerz.cc/resources)

## Links

- [Full API & CLI Reference](https://layerz.cc/llms-full.txt)
- [Agent Integration Guide](https://layerz.cc/for-agents)
- [Get Started](https://layerz.cc)