D365 F&O MCP Server
README.md
# D365 F&O MCP Server
<div align="center">
**22 AI tools that know every X++ class, table, form, and EDT in your D365FO codebase**
[](https://www.npmjs.com/package/d365fo-mcp)
[](https://opensource.org/licenses/MIT)
[](https://nodejs.org/)
[](https://www.typescriptlang.org/)
[](docs/TESTING.md)
<!-- coverage-badge:start -->
[](eval/COVERAGE.md) [](eval/COVERAGE.md)
<!-- coverage-badge:end -->
*Grounded AI development for Dynamics 365 Finance & Operations — works with GitHub Copilot and Claude Code*
[](https://insiders.vscode.dev/redirect/mcp/install?name=d365fo&inputs=%5B%7B%22type%22%3A%22promptString%22%2C%22id%22%3A%22d365fo_server_url%22%2C%22description%22%3A%22D365FO%20MCP%20server%20URL%20(e.g.%20https%3A%2F%2Fyour-server.azurewebsites.net%2Fmcp%2F)%22%7D%5D&config=%7B%22type%22%3A%22http%22%2C%22url%22%3A%22%24%7Binput%3Ad365fo_server_url%7D%22%7D)
[](https://insiders.vscode.dev/redirect/mcp/install?name=d365fo&quality=insiders&inputs=%5B%7B%22type%22%3A%22promptString%22%2C%22id%22%3A%22d365fo_server_url%22%2C%22description%22%3A%22D365FO%20MCP%20server%20URL%20(e.g.%20https%3A%2F%2Fyour-server.azurewebsites.net%2Fmcp%2F)%22%7D%5D&config=%7B%22type%22%3A%22http%22%2C%22url%22%3A%22%24%7Binput%3Ad365fo_server_url%7D%22%7D)
[](https://cursor.com/install-mcp?name=d365fo&config=eyJ1cmwiOiJodHRwczovL3lvdXItc2VydmVyLmF6dXJld2Vic2l0ZXMubmV0L21jcC8ifQ%3D%3D)
*These connect an editor to a server that is already deployed — see [Quick Start](#quick-start) if you still need to set one up.*
</div>
---
## Why
AI assistants excel at C#, Python, and JavaScript. X++ is different: your D365FO codebase is private, deeply customized, and invisible to every model — so AI confidently generates code that doesn't compile.
This server pre-indexes your entire D365FO installation (580 000+ symbols across standard, ISV, and custom models) and exposes it as 22 specialized MCP tools (`axdb_sql` and `trace_d365fo` only once configured). Every signature, every CoC wrapper, every label, every form pattern — verified against your real metadata **before** the AI writes a single line.
| Task | Without this server | With this server |
|------|--------------------|------------------|
| Method signatures | Guessed → compile errors | Exact, from your codebase |
| Existing CoC wrappers | Manual AOT search | `extension_info(mode="coc")` in < 50 ms |
| New forms | Hand-written XML, broken patterns | Cloned from reference forms, validated against the pattern catalog |
| Labels | Hardcoded strings | Right `@SYS`/`@MODULE` key found instantly |
| Security chains | Hours of manual tracing | Role → Duty → Privilege → Entry Point in one call |
| Generated code | Hallucinated fields and types | Every reference proven against the index, gated before write |
---
## Capabilities
| Feature | Description |
|---|---|
| 🔍 **Full-codebase intelligence** | 580K+ symbols indexed: classes, tables, forms, EDTs, enums, labels (20M+ rows), security artifacts — FTS5 search in < 10 ms |
| 🛡️ **Grounded generation** | Fail-closed gates: `prepare` issues grounding tokens, `validate_code(mode="references")` proves every identifier, `validate_code(mode="syntax")` enforces best practices — hallucinated code never reaches disk |
| 🧩 **Form pattern engine** | Complete catalog of Microsoft form patterns and sub-patterns: recommends the right pattern, clones reference forms with datasource re-binding, **deterministically expands** patterns that have no reference form, **auto-repairs** a form's missing required controls, validates structure and blocks invalid writes |
| ✍️ **Safe metadata writes** | C# bridge uses Microsoft's own `IMetadataProvider` wherever it can express the object; the few types and ops it cannot go through structured XML writers with ambiguity guards — never blind string replacement. Automatic `.rnrproj` registration, one-call undo |
| 🏗️ **SDLC integration** | MSBuild compilation with structured diagnostics, DB sync, xppbp best practices, SysTestRunner — all from chat. Optional `restartAos: true` refreshes the matching local IIS/IIS Express runtime after a successful build and requested sync. [Runtime restart](docs/AOS_RUNTIME_RESTART.md) |
| 🗄️ **Optional AxDB SQL** | Off by default and not even published until configured: `axdb_sql` runs read-only, parameterized queries against AxDB on the development VM, so the agent can check what the code actually persisted. Windows authentication; one SELECT checked on the parsed T-SQL tree, inside a transaction that is always rolled back; no result cache. [Setup](docs/AXDB_SQL.md) |
| ⏱️ **Optional performance trace** | Off by default and not published until enabled: `trace_d365fo` captures the AOS's ETW events (X++ method enter/exit, SQL latency) on the development VM while the agent — or you — reproduce a slow scenario, and analyses the capture: where the time goes per method, which statements run thousands of times from one caller. No timer code added to the application; parameter values never returned. [Setup and accuracy](docs/TRACE.md) |
| 📐 **X++ knowledge base** | Queryable rules: select grammar, CoC authoring, financial dimensions, the posting engine (`LedgerVoucher`), number sequences, `SysExtension`, Electronic Reporting, AX2012→D365FO migration — prevents deprecated APIs |
### Pattern-grounded form development
Forms are the hardest artifact to generate correctly — each pattern dictates required containers, ordering, and allowed sub-patterns. The form pattern engine makes it a guided pipeline:
```mermaid
flowchart LR
A["object_patterns<br/>(domain=form, action=analyze)"] --> B["object_patterns<br/>(domain=form, action=spec)"]
B --> C["generate_object<br/>objectType=form, cloneFrom"]
C --> D["object_patterns<br/>(domain=form, action=validate) FP001–FP010"]
D -->|clean| E["d365fo_file<br/>(action=create) write + project"]
D -->|errors| C
```
Structural violations (wrong order, missing container, disallowed control) **block the write** — recommendations only warn. Mined pattern statistics from your own environment ground every suggestion in reality.
---
## Quick Start
> **From D365FO platform update 10.0.49 (PU74), Visual Studio 2026 is the supported IDE for X++ development** — Microsoft no longer supports VS 2022. Earlier platform versions still use VS 2022 ≥ 17.14. [Details](https://learn.microsoft.com/en-us/dynamics365/fin-ops-core/fin-ops/get-started/whats-new-platform-updates-10-0-49)
**Installing on your own D365FO VM** — the usual case. One line in PowerShell installs Node.js if it is missing, installs the server from npm, and runs the setup wizard, which asks where the index should live and builds the C# bridge for you:
```powershell
irm https://raw.githubusercontent.com/dynamics365ninja/d365fo-mcp-server/main/install.ps1 | iex
```
Already have Node.js 24+? Then the one-liner has nothing to bootstrap and you can skip it:
```powershell
npm install -g d365fo-mcp
d365fo-mcp setup
```
Re-running either is safe. An installation made before the npm package existed is a git checkout, and both are left exactly where they are and updated in place.
**Your team already runs a shared server?** Then you install nothing — point your editor at it:
```powershell
npx d365fo-mcp connect https://your-server.azurewebsites.net
```
Both paths in full — prerequisites, editor configuration for every scenario, the required instruction file, and how to verify grounding actually works: **[docs/QUICK_START.md](docs/QUICK_START.md)**
---
## Azure Deployment
One shared instance for the whole team — the metadata index lives in Blob Storage and downloads automatically on startup.
[](https://portal.azure.com/#create/Microsoft.Template/uri/https%3A%2F%2Fraw.githubusercontent.com%2Fdynamics365ninja%2Fd365fo-mcp-server%2Frefs%2Fheads%2Fmain%2Finfrastructure%2Fazuredeploy.json)
Deployment guide: [docs/SETUP_AZURE.md](docs/SETUP_AZURE.md) — includes CI/CD pipeline automation
---
## Documentation
| Getting started | Reference | Operations |
|-----------------|-----------|------------|
| [Quick Start](docs/QUICK_START.md) — connect or install | [All 22 tools](docs/MCP_TOOLS.md) | [Azure deployment](docs/SETUP_AZURE.md) |
| [Setup scenarios A–F](docs/SETUP.md) | [`.mcp.json` reference](docs/MCP_CONFIG.md) | [DevOps pipelines](docs/SETUP_AZURE.md#azure-devops-pipelines) |
| [Claude Code setup](docs/SETUP.md#claude-code-cli) | [Configuration](docs/CONFIGURATION.md) | [Testing](docs/TESTING.md) |
| [Usage examples](docs/USAGE_EXAMPLES.md) — real tool chains | [Architecture](docs/ARCHITECTURE.md) | [Custom / ISV models](docs/CUSTOM_EXTENSIONS.md) |
| [Changelog](CHANGELOG.md) | [Knowledge authoring](docs/KNOWLEDGE_AUTHORING.md) | [Coverage](eval/COVERAGE.md) — what the badge counts |
| [Backlog](docs/BACKLOG.md) — deferred work | [New tool checklist](docs/NEW_TOOL_CHECKLIST.md) | [Eval loop](docs/AGENT_EVAL_LOOP.md) — the self-improvement harness |
## License
MIT
TDQS
A4.1/5.0
Scored across 20 tools
Disambiguation4/5
Most tools target distinct resources/actions, and the descriptions actively disambiguate (e.g. 'use X instead of Y', 'already folded into...'). However, genuine overlaps exist: `prepare` deliberately duplicates search/get_object_info/analyze_code/generate_object, `d365fo_file(action="generate")` overlaps `generate_object`, and build/run_bp_check/verify_d365fo_project form a pipeline where invocation boundaries must be inferred. These are explained but still create selection risk.
Naming Consistency4/5
Casing is uniformly snake_case and most names are clear verb_noun (get_object_info, run_bp_check, generate_object, validate_code, find_references). A minority are bare nouns/vague verbs (prepare, search, labels, object_patterns, security_info, extension_info, d365fo_file), which is a minor inconsistency but still readable and predictable.
Tool Count4/5
20 tools is above the ideal 3-15 band and trends heavy, but the domain (D365 F&O code generation, indexing, builds, validation, security, patterns, knowledge) is genuinely broad, so each tool maps to a distinct capability. It sits slightly over-scoped rather than padded.
Completeness4/5
The surface covers an unusually full authoring lifecycle: context discovery, generation, validation, naming, indexing, labels, build/verify, extensibility analysis, security, and knowledge lookup. Minor gaps remain around deployment/package, data management, and environment operations, but core CRUD/lifecycle workflows for X++/AOT authoring are covered with no obvious dead ends.
Maintenance
ActivityActive
ResponsivenessResponsive