Accounting & Finance
Settlements, P&L, journals, COGS, and reconciliation as MCP tools on the Financials suite.
MCP Access
Ask your AI about inventory, profit, and operations
Connect ChatGPT, Claude, Cursor, or any MCP-compatible assistant to NeonPanel. Ask in plain English — get answers from live inventory, financials, and analytics instead of exporting reports
- Accounting
- Inventory
- Forecasting
- AI access
Trusted by 200+ ecommerce teams
What is MCP Access?
A direct line from your AI to NeonPanel
MCP Access is AI access to ecommerce data through an ecommerce MCP server: AI assistants connect to your NeonPanel account through MCP (Model Context Protocol) — often described as a universal plug for AI tools.
Instead of opening reports, exporting to Excel, and analyzing manually, you ask a question in chat. Your assistant pulls live data from NeonPanel and returns the answer.
What can you ask?
Example questions teams ask every day
Most teams start with operational questions they would otherwise answer across several dashboards.
Inventory
- Which SKUs will run out in the next 30 days?
- What inventory should I reorder this week?
- Show products with excess stock.
Finance
- What was my net profit last month?
- Compare profitability by brand.
- Why did margins decrease?
Operations
- Show outstanding supplier payments.
- Which shipments are delayed?
- Summarize this week’s operational risks.
Supported AI tools
Works with the assistants you already use
MCP Access is an integration layer — not a separate NeonPanel feature locked to one chatbot. Any client that supports custom MCP connections can connect.
Claude
ChatGPT
Cursor
Windsurf
VS Code
Gemini
Custom MCP clients
For structured NeonPanel agents without wiring MCP yourself, see NeonPanel for Claude and Agents in Slack.
How it works
Three steps to live answers
1
Connect your AI assistant
Add NeonPanel MCP in Claude, Cursor, ChatGPT, or your preferred MCP-compatible client.
2
Authorize NeonPanel
Sign in with OAuth and approve access for your company — scoped to your role.
3
Ask and get insights
Questions run against live ledger data; answers arrive in the conversation.
Without MCP
NeonPanel → export report → Excel → manual analysis
With MCP Access
NeonPanel → MCP Access → your AI — answer in chat
Why MCP Access
Why use MCP instead of reports?
Same NeonPanel ledger — fewer exports, filters, and spreadsheet rebuilds before you get an answer.
| Traditional reporting | MCP Access |
|---|---|
| Open multiple reports | Ask one question |
| Export to Excel | Instant answer in chat |
| Build filters manually | Natural language |
| Analyze data yourself | AI analyzes live NeonPanel data |
Security & permissions
Built for finance-grade data
CFOs and operators need to know AI access is controlled. MCP Access uses OAuth authorization, role-based NeonPanel permissions, and revocable connections.
Which tools an assistant can call depends on the client surface (Claude, Slack, NeonaSphera, or a raw MCP IDE) and your plan — not one global setting across every channel.
- OAuth authorization
- Role-based permissions
- Surface-specific tool access
- Data never shared publicly
- Revoke access anytime
Who benefits
Who MCP Access is for
Operators, finance, agencies, and AI-powered teams ask the same ledger — without rebuilding reports for each role.
Ecommerce owners & operations managers
Ecommerce owners stop digging through reports and get instant answers about sales, inventory, profitability, and growth opportunities. Operations managers monitor inventory, purchase orders, and supply chain performance in natural language.
Finance teams & CFOs
Finance teams & CFOs analyze profitability, cash flow, margins, and financial performance without manually building reports.
Agencies, consultants & AI-powered teams
Agencies & consultants access client data faster and generate insights without hours in dashboards. AI-powered teams connect ChatGPT, Claude, Cursor, or custom MCP clients directly to business data and automate analysis workflows.
Where this sits · For technical teams
Where MCP sits in the stack
AI on the ledger Accounting, Inventory, and Forecasting already use — structured read and write against the same numbers that close the books. Structured tool responses on the reconciled ledger: reconciled objects and tool payloads — not CSV dumps — so models reason on NeonPanel’s ledger, not reconstructed spreadsheets.
Inventory Management
Locations, batches, landed costs, valuation, and inbound shipments as queryable Inventory tools.
Forecasting & Replenishment
Forecast scenarios, velocity, lead times, and stockout risk when Forecasting is enabled on your plan.
Three suites your model can call
Brand Analytics, Financials, and Inventory — plus Forecasting when enabled on your plan.
Brand Analytics
Financials
Inventory
Brand Analytics
Brand Analytics as MCP tools: query catalog and search performance, then join to your NeonPanel financials in the same session without re-exporting ranges.
- Search Catalogue Performance — impressions through purchases, conversion rates per ASIN
- Search Query Performance — which search terms drive your brand, query volume, click/cart/purchase funnels
- Top Search Terms — category-level keyword demand, top 3 products and brands per term
- Customer Loyalty Analytics — repeat purchase trends, lifetime value segmentation
- Repeat Purchase Behaviour — new vs. returning customer splits
- Market Basket Analysis — co-purchased products, cross-sell opportunities
- Demographics — age, income, education, gender breakdowns (US market)
Financials
Settlements, fees, refunds, FIFO COGS, and journals as structured tool payloads — the accounting-grade slice of the ledger your close already depends on.
- P&L — by SKU, by entity, consolidated
- COGS — with full landed cost breakdown, FIFO-calculated
- Margin analysis — revenue trends, SKU-level profitability
- Journal entries — accounting system health and audit trail
- Settlement reconciliation — marketplace fees, refunds, adjustments
Inventory
FBA, AWD, 3PL, and internal warehouses in one queryable state — transfers, adjustments, and inbound receipts with traceability.
- Stock positions — across FBA, AWD, 3PL, and internal warehouses simultaneously
- Transfer and adjustment history — with reason codes and audit traceability
- Inbound shipment tracking — receipt events, batch arrivals, processing status
- Multi-warehouse movements — full inventory transaction history across all locations
Also available via MCP
Forecasting & Replenishment on the same ledger: velocity, reorder guidance, stockout risk, and replenishment scenarios — same MCP interface when Forecasting is enabled.
Forecasting module required. Available on Professional and Enterprise plans.
See the shape of the data
A single MCP tool call. A structured response your AI can reason over immediately.
Tool call
get_sku_margin_with_search_performance(
period: "2026-03",
top_n: 5,
sort_by: "gross_margin",
include: [
"search_query_performance",
"cogs_breakdown"
]
)Response
[
{
sku: "NP-OMEGA-12",
revenue: 48200,
cogs: 19800,
gross_margin: 0.589,
landed_cost_per_unit: 4.12,
top_search_terms: [
"omega supplement",
"fish oil 1000mg"
],
search_query_trend: "↑ +18% MoM click share",
purchase_funnel: {
impressions: 142000,
clicks: 3800,
purchases: 610
}
},
...
]Structured fields from live NeonPanel state — ready for downstream reasoning without cleanup passes.
Live in Slack
Agents in Slack
@mention Dorian, Marta, Nina, and the rest of the default NeonPanel roster directly — same specialists as Claude, without wiring MCP yourself.
Interactive by @mention with no schedule controls in Slack.
Also available
Viktor in Slack & Teams
Prefer one general coworker? Viktor routes questions to the same MCP tools — useful when your team does not want to pick a specialist. Optional router, not required for Agents in Slack.
Same ledger in chat
Ask in Slack, get structured NeonPanel answers
Specialist agents call NeonPanel MCP tools on your live ledger and reply in-thread — inventory, settlements, and forecasting without leaving the channel.
Works with Agents in Slack and Viktor as an optional router. No schedule controls in Slack.
Need multi-step agent fleets with routing and guardrails beyond raw MCP? See NeonaSphera — NeonPanel for Claude → · See pricing →
Frequently asked questions
MCP Access connects AI assistants to your NeonPanel account through Model Context Protocol (MCP). You ask questions in plain language and get answers based on live inventory, financial, and operational data.
Any client that supports custom MCP connections — including Claude, ChatGPT, Cursor, Windsurf, VS Code, and Gemini. For pre-packaged NeonPanel agents, see NeonPanel for Claude and Agents in Slack.
Inventory levels and reorder timing, profitability and margins, supplier payments, shipment status, forecasting scenarios, and brand analytics — anything exposed as MCP tools on your plan.
Reporting requires opening dashboards, exporting files, and analyzing manually. MCP Access lets you ask one question in chat and receive an answer from live NeonPanel data.
Not for agent surfaces like Claude or Slack — you ask in plain English. Connecting a raw MCP client in an IDE takes basic setup steps documented for developers.
Yes. Connections use OAuth authorization, role-based NeonPanel permissions, and can be revoked at any time. Your data stays in your NeonPanel tenant.
Available actions depend on which surface you use (Claude, Slack, NeonaSphera, or a direct MCP client) and your NeonPanel role — not one global setting. Analysis is always available; write tools follow your plan and permissions.
Only data in your authorized NeonPanel company — inventory, settlements, COGS, forecasting, brand analytics, and related operational records scoped to your role.
Most teams connect an assistant and authorize NeonPanel in a few minutes. OAuth handles authentication on first use.
Yes. Revoke the connection from NeonPanel or your AI client workspace whenever you need to.
Ready to connect?
Available as a company-wide add-on on any plan. Book a demo or start free — See pricing for packaging.