Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@rb209-mcp What's the RB209 recommendation for winter wheat?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server exposing the AHDB RB209 (Nutrient Management Guide) Web API v2 — 94 tools — plus a nutrient-planning skill.
The three Users auth endpoints (Login / Logout / Refresh_Token) are deliberately
not exposed: the server manages that token lifecycle itself, so a tool call to
Logout could only ever invalidate its own session.
The extension ships its own usage guidance, so installing the .mcpb in Claude Desktop is enough — you don't need to load the skill separately. The server exposes an MCP instructions brief on connect, and the highest-risk tools (notably the recommendation endpoint) carry the must-know rules in their tool descriptions. skill/rb209-nutrient-planning/SKILL.md remains the fuller guide (for Claude Code / claude.ai and as a versioned release asset); the embedded guidance is a condensed subset kept in sync with it by hand.
Tools are a faithful passthrough to the API with one deliberate exception: the
recommendation endpoint has a footgun where a missing field.grass (which the
API's validation does not report) makes the call fail with the misleading
422 "…calculating the crop order". For that endpoint only, the server fills
the incidental required-empty sections (grass, grassland, organicMaterials,
mannerOutputs) when they are absent — never overwriting supplied data, and with
no effect on the result. See src/recommendationGuardrail.ts.
You need your own AHDB RB209 Web API licence (register at https://rb209.ahdb.org.uk/Home/RequestAccess). Each user supplies their own credentials; they never leave your machine.
Related MCP server: nutrition-mcp
Download rb209-mcp.mcpb from the latest Release.
Double-click it; Claude Desktop opens the installer.
Enter your RB209 email and password in the settings form.
Done — the RB209 tools appear in your chat tool.
npm install && npm run build
RB209_EMAIL=you@example.com RB209_PASSWORD=secret node dist/index.jsAdd to an MCP client config with command: node, args: [".../dist/index.js"],
and env: { RB209_EMAIL, RB209_PASSWORD }.
Env var | Required | Default | Purpose |
| yes | — | Your RB209 licence email |
| yes | — | Your RB209 licence password |
| no |
| API base URL |
| no | all | Comma-separated tags to expose (e.g. |
| no | vendored spec | Only set when running a manually-extracted bundle standalone (the .mcpb manifest wires it automatically) |
npm test # unit tests (msw-mocked)
npm run test:watch
RB209_EMAIL=... RB209_PASSWORD=... npm test # also runs the live integration testThis server is released under the MIT Licence. The RB209 data it returns is licensed API content © AHDB under the Open Government Licence v3.0.
GET /api/AdviceNote/AdviceNote/{adviceNoteCode} — Individual advice note text filtered from the supplied corresponding advice note code.
Inputs: adviceNoteCode (required).
GET /api/AdviceNote/AdviceNotes — The full list of available advice notes.
Inputs: no named parameters.
GET /api/AdviceNote/ArableEngWal — The full list of available advice notes for England and Wales arable.
Inputs: no named parameters.
GET /api/AdviceNote/ArableScot — The full list of available advice notes for Scotland arable.
Inputs: no named parameters.
GET /api/AdviceNote/GrassEngWal — The full list of available advice notes for England and Wales grass.
Inputs: no named parameters.
GET /api/AdviceNote/GrassScot — The full list of available advice notes for Scotland grass.
Inputs: no named parameters.
GET /api/AdviceNote/AdviceNoteText/{countryId}/{adviceNoteCode} — Individual advice note text filtered from the supplied country id and advice note code.
Inputs: countryId (required), adviceNoteCode (required).
GET /api/Arable/CropGroup/{cropGroupId} — Individual crop group - filtered by crop group id.
Inputs: cropGroupId (required).
GET /api/Arable/CropGroups — The full list of available crop groups.
Inputs: no named parameters.
GET /api/Arable/CropInfo1/{cropTypeId}/{cropInfo1Id} — Individual crop info 1 - filtered by crop type id and crop info 1 id.
Inputs: cropTypeId (required), cropInfo1Id (required).
GET /api/Arable/CropInfo1s — The full list of available crop info 1s.
Inputs: no named parameters.
GET /api/Arable/CropInfo1s/{cropTypeId} — A filtered list of available crop info 1s - filtered by crop type id.
Inputs: cropTypeId (required).
GET /api/Arable/CropInfo2/{cropInfo2Id} — Individual crop info 2 - filtered by crop info 2 id (only required for arable cereals crops).
Inputs: cropInfo2Id (required).
GET /api/Arable/CropInfo2s — The full list of available crop info 2s (only required for arable cereals crops).
Inputs: no named parameters.
GET /api/Arable/CropType/{cropTypeId} — Individual crop type - filtered by crop type id.
Inputs: cropTypeId (required).
GET /api/Arable/CropTypes — The full list of available crop types.
Inputs: no named parameters.
GET /api/Arable/CropTypes/{cropGroupId} — A filtered list of available crop types - filtered by crop group id.
Inputs: cropGroupId (required).
GET /api/Arable/DefaultYields/{countryId} — Get the default yield data for the crops.
Inputs: countryId (required).
GET /api/Arable/PotatoGroup/{potatoGroupId} — Individual potato group - filtered by potato group id (only required for arable potato crops).
Inputs: potatoGroupId (required).
GET /api/Arable/PotatoGroups — The full list of available potato groups (only required for arable potato crops).
Inputs: no named parameters.
GET /api/Arable/PotatoVarieties — The full list of available potato varieties (only required for arable potato crops).
Inputs: no named parameters.
GET /api/Arable/PotatoVarieties/{potatoGroupId} — A filtered list of available potato varieties - filtered by potato group id (only required for arable potato crops).
Inputs: potatoGroupId (required).
GET /api/Arable/PotatoVariety/{potatoVarietyId} — Individual potato variety - filtered by potato variety id (only required for arable potato crops).
Inputs: potatoVarietyId (required).
GET /api/FertiliserPrices/LatestPrices — Get the latest prices of fertilisers.
Inputs: no named parameters.
POST /api/FertiliserPrices/RequestAccess — Request access to the fertiliser latest prices endpoint.
Inputs: body (required).
GET /api/Field/Countries — The full list of available countries.
Inputs: no named parameters.
GET /api/Field/Country/{countryId} — Individual country text filtered from the supplied corresponding country id.
Inputs: countryId (required).
GET /api/Field/Nutrient/{nutrientId} — Individual nutrient text filtered from the supplied corresponding nutrient id.
Inputs: nutrientId (required).
GET /api/Field/Nutrients — The full list of available nutrients.
Inputs: no named parameters.
GET /api/Field/SecondCropType_List/{cropGroupId1}/{cropTypeId1}/{cropGroupId2}/{countryId} — Some crops (both arable and grassland) allow a second crop to be sown in the same crop year: The filtered list of available second crop types after the first crop.
Inputs: countryId (required), cropTypeId1 (required), cropGroupId1 (required), cropGroupId2 (required).
GET /api/Field/SiteClass/{siteClassId} — Site class of site class id provided.
Inputs: siteClassId (required).
GET /api/Field/SiteClasses/{countryId} — Site class list.
Inputs: countryId (required).
GET /api/Field/SiteClassItem/{soilTypeId}/{altitude}/{postcode}/{countryId} — Site class of field.
Inputs: altitude (required), postcode (required), countryId (required), soilTypeId (required).
GET /api/Field/FieldType/{fieldTypeId} — Individual field type text filtered from the supplied corresponding field type id.
Inputs: fieldTypeId (required).
GET /api/Field/FieldTypes/{countryId} — The full list of available field types. — You usually don't need this — fieldType is fixed: 1 = Arable & Horticulture, 2 = Grassland, 3 = Both. It only returns data for countryId 3.
Inputs: countryId (required).
GET /api/Grass/DefoliationSequence/{defoliationSequenceId} — Get the text of the defoliation sequence for the provided defoliation sequence id.
Inputs: defoliationSequenceId (required).
GET /api/Grass/DefoliationSequences/{swardTypeId}/{swardManagementId}/{numberOfCuts}/{newSward} — Deprecated - The list of different defoliation sequences available for grass fields.
Inputs: newSward (required), swardTypeId (required), numberOfCuts (required), swardManagementId (required).
GET /api/Grass/DefoliationSequences/{swardTypeId}/{swardManagementId}/{numberOfCuts}/{newSward}/{countryId} — The list of different defoliation sequences available for grass fields.
Inputs: newSward (required), countryId (required), swardTypeId (required), numberOfCuts (required), +1 more.
GET /api/Grass/GrassGrowthClass/{grassGrowthClassId} — Individual grass growth class (GGC) text filtered from the supplied corresponding grass growth class id.
Inputs: grassGrowthClassId (required).
GET /api/Grass/GrassGrowthClass/{soilTypeId}/{rainfall}/{altitude}/{chalk} — Calculate the grass growth class (GGC) for a grassland field.
Inputs: chalk (required), altitude (required), rainfall (required), soilTypeId (required).
GET /api/Grass/GrassGrowthClasses/{countryId} — Full list of available grass growth classes (GGC) for grassland fields.
Inputs: countryId (required).
GET /api/Grass/GrassHistories/{countryId} — Grass history list of grass fields.
Inputs: countryId (required).
GET /api/Grass/GrassHistory/{grassHistoryId} — Grass history of grass history id provided.
Inputs: grassHistoryId (required).
GET /api/Grass/PotentialCuts/{swardTypeId}/{swardManagementId} — The list of different number of cuts for the field.
Inputs: swardTypeId (required), swardManagementId (required).
GET /api/Grass/GrassSeason/{seasonId} — This endpoint is used to return grass season.
Inputs: seasonId (required).
GET /api/Grass/GrassSeasons/{countryId} — This endpoint is used to return grass seasons.
Inputs: countryId (required).
GET /api/Grass/SiteClassId/{soilTypeId}/{rainfall}/{altitude} — Calculate the site class id for Scottish grassland fields.
Inputs: altitude (required), rainfall (required), soilTypeId (required).
GET /api/Grass/SwardManagement/{swardManagementId} — Get the text of the sward management for the provided sward management id.
Inputs: swardManagementId (required).
GET /api/Grass/SwardManagements — The list of different sward managements available for grass fields.
Inputs: no named parameters.
GET /api/Grass/SwardManagements/{swardTypeId} — The list of different sward managements available for the provided sward type.
Inputs: swardTypeId (required).
GET /api/Grass/SwardType/{swardTypeId} — Get the text of the sward type for the provided sward type id.
Inputs: swardTypeId (required).
GET /api/Grass/SwardTypes — The list of different sward types available for grass fields.
Inputs: no named parameters.
GET /api/Grass/YieldRangesEnglandAndWales/{sequenceId}/{grassGrowthClassId} — The list of yield ranges for grass fields in England and Wales.
Inputs: sequenceId (required), grassGrowthClassId (required).
GET /api/Grass/YieldRangesScotland/{sequenceId}/{siteClassId} — The list of yield ranges for grass fields in Scotland.
Inputs: sequenceId (required), siteClassId (required).
GET /api/Measurement/CropHeights — Full list of available crop heights.
Inputs: no named parameters.
GET /api/Measurement/GreenAreaIndexes — Full list of available green area indexes.
Inputs: no named parameters.
POST /api/Measurement/MeasurementMethod — The endpoint to calculate sns index using the measurement method in England & Wales.
Inputs: body (required).
POST /api/MeasurementScotland/MeasurementMethod — The endpoint to calculate residue group using SMN measurement values in Scotland.
Inputs: body (required).
GET /api/Measurement/Seasons — Full list of available seasons.
Inputs: no named parameters.
GET /api/Measurement/ShootNumbers — Full list of available shoot numbers.
Inputs: no named parameters.
GET /api/Measurement/SmnConversionMethod/{smnValue}/{soilLayer} — The get smn value to be converted from N/kg to kg/ha.
Inputs: smnValue (required), soilLayer (required).
GET /api/OrganicMaterial/OrganicMaterialCategories — This endpoint is used to return organic material categories.
Inputs: no named parameters.
GET /api/OrganicMaterial/OrganicMaterialCategoryItem/{organicMaterialCategoryId} — This endpoint is used to return organic material category item.
Inputs: organicMaterialCategoryId (required).
GET /api/OrganicMaterial/IncorporationMethod/{incorporationMethodId} — Organic material incorporation method item based on incorporation method id.
Inputs: incorporationMethodId (required).
GET /api/OrganicMaterial/IncorporationMethods — Organic material incorporation method list.
Inputs: no named parameters.
GET /api/OrganicMaterial/IncorporationMethods/{organicMaterialTypeId} — Organic material incorporation method list based on organicMaterialTypeId.
Inputs: organicMaterialTypeId (required).
GET /api/OrganicMaterial/OrganicMaterialTypeItem/{organicMaterialTypeId} — This endpoint is used to return organic material type of organic material type id provided.
Inputs: organicMaterialTypeId (required).
GET /api/OrganicMaterial/OrganicMaterialTypeItem/{organicMaterialTypeId}/{dryMatterSplit} — This endpoint is used to return organic material type item for provided organicMaterialTypeId and dryMatterSplit.
Inputs: dryMatterSplit (required), organicMaterialTypeId (required).
GET /api/OrganicMaterial/OrganicMaterialTypes — This endpoint used to return organic material types.
Inputs: no named parameters.
GET /api/OrganicMaterial/OrganicMaterialTypes/{organicMaterialCategoryId} — This endpoint used to return organic material types for provided organicMaterialCategoryId.
Inputs: organicMaterialCategoryId (required).
GET /api/OrganicMaterial/OrganicMaterialTypes/{organicMaterialCategoryId}/{dryMatterSplit} — This endpoint used to return organic material types for provided organicMaterialCategoryId and dryMatterSplit.
Inputs: dryMatterSplit (required), organicMaterialCategoryId (required).
GET /api/OrganicMaterial/OrganicMaterialTypes/{dryMatterSplit} — This endpoint used to return organic material types based on filter.
Inputs: dryMatterSplit (required).
GET /api/PreviousCropping/PreviousGrass/{previousGrassId} — Individual previous grass - filtered by previous grass id.
Inputs: previousGrassId (required).
GET /api/PreviousCropping/PreviousGrasses — The full list of available previous grasses.
Inputs: no named parameters.
GET /api/Rainfall/RainfallAverage/{postcode} — Average rainfall of the field across the whole year.
Inputs: postcode (required).
GET /api/Rainfall/SummerRainfallAverage/{postcode} — Average rainfall of the field during the summer months (April - September).
Inputs: postcode (required).
GET /api/Rainfall/WinterRainfallAverage/{postcode} — Average rainfall of the field during the winter months (October - March).
Inputs: postcode (required).
GET /api/Recommendation/CalculateNutrientDeficiency/{cropTypeId}/{leafSamplingPosition}/{nutrientId}/{nutrientContent} — To get the nutrient deficiency result bases on leaf analysis.
Inputs: cropTypeId (required), nutrientId (required), nutrientContent (required), leafSamplingPosition (required).
POST /api/Recommendation/CalculateNutrientOfftake — The calculate crop nutrient offtake value.
Inputs: body (required).
POST /api/Recommendation/Recommendations — The main connection to calculate nutrient recommendations. — Before calling, collect ALL inputs from the user — do not assume rainfall, sowing date, yield or straw handling; get rainfall from a postcode via rb209_rainfall_average. Call this ONCE — it returns every nutrient you enable in nutrients (N, P2O5, K2O, MgO, SO3, lime); do not call it per-nutrient or repeat it. For an ARABLE field the body MUST include field.grass = {} AND field.grassland = {} (both empty objects) alongside field.arable = [ ... ]; a missing field.grass causes the misleading 422 '…calculating the crop order', not a 'grass required' error (this server auto-fills them, but be explicit). No soil analysis? send field.soil.soilAnalyses = [] and skip the soil index/methodology lookups. cropInfo2Id (1 = straw baled/removed, 2 = straw incorporated) strongly affects K — always ask the user. Resolve all IDs from the lookup tools.
Inputs: body (required).
GET /api/Soil/Methodologies/{nutrientId}/{countryId} — A filtered list of available soil methodologies - filtered by nutrient id and country id.
Inputs: countryId (required), nutrientId (required).
GET /api/Soil/Methodology/{nutrientId}/{methodologyId} — Individual soil methodology - filtered by nutrient id and methodology id.
Inputs: nutrientId (required), methodologyId (required).
GET /api/Soil/NutrientIndex/{nutrientId}/{indexId} — Individual nutrient index item - filtered by nutrient id and index id.
Inputs: indexId (required), nutrientId (required).
GET /api/Soil/NutrientIndex/{nutrientId}/{nutrientValue}/{methodologyId} — Individual nutrient index item - filtered by nutrient id, nutrient value and methodology id.
Inputs: nutrientId (required), methodologyId (required), nutrientValue (required).
GET /api/Soil/NutrientIndex/{nutrientId}/{nutrientValue}/{methodologyId}/{countryId} — Individual nutrient index item - filtered by nutrient id, nutrient value, methodology id and country Id.
Inputs: countryId (required), nutrientId (required), methodologyId (required), nutrientValue (required).
GET /api/Soil/NutrientIndexes/{methodologyId} — A list of the available nutrient indexes across nutrients - filtered by methodology id. — Returns T04 'No nutrient indexes found' for many nutrient/methodology combos. If the user has no soil analysis, skip it and send soilAnalyses: []; only convert a measured value via the NutrientIndexIdFromValue tool.
Inputs: methodologyId (required), nutrientId.
GET /api/Soil/NutrientIndexIdFromValue/{nutrientId}/{methodologyId}/{nutrientValue}/{countryId} — Individual nutrient index id - filtered by nutrient id, methodology id and nutrient value.
Inputs: countryId (required), nutrientId (required), methodologyId (required), nutrientValue (required).
GET /api/Soil/NutrientIndices/{nutrientId}/{methodologyId}/{countryId} — A filtered list of available nutrient indexes - filtered by nutrient id, methodology id and country id.
Inputs: countryId (required), nutrientId (required), methodologyId (required).
GET /api/Soil/NutrientTargetIndex/{cropGroupId}/{nutrientId}/{countryId} — Get the nutrient target index.
Inputs: countryId (required), nutrientId (required), cropGroupId (required).
GET /api/Soil/NvzActionProgram/{countryId} — The full list of NVZ action program.
Inputs: countryId (required).
GET /api/Soil/SoilPhRecommendation/{soilTypeId}/{countryId} — Get list of soil ph recommendation level.
Inputs: countryId (required), soilTypeId (required).
GET /api/Soil/SoilPsc/{cropGroupId}/{pIndexId} — Get the list of all available PSC for selected crop group.
Inputs: pIndexId (required), cropGroupId (required).
GET /api/Soil/SoilType/{soilTypeId} — Individual soil type - filtered by soil type id.
Inputs: soilTypeId (required).
GET /api/Soil/SoilTypes — The full list of available soil types.
Inputs: no named parameters.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 94 tools
The set has many near-identical lookup families—e.g., at least four soil nutrient-index-by-value variants and multiple organic-material-type filters—so an agent must parse subtle parameter differences to avoid misselection. Singular/plural pairs and a deprecated endpoint duplicated by a country-aware version further blur boundaries.
All tools share a uniform `rb209_<area>_<resource>` snake_case pattern, which is predictable. However, naming is inconsistent at the resource level: `soil_nutrient_indices` vs `soil_nutrient_indexes` are both present, singular and plural forms are mixed, and some names are awkwardly long (`rb209_grass_growth_class_by_soil_type_rainfall_altitude_chalk`).
94 tools is an extreme tool count for an MCP server, far beyond the point where an agent can reliably choose. Many endpoints are trivial singular/plural pairs or filter variants that could be collapsed into parameterized tools, so this feels like a direct API surface dump rather than a curated tool set.
For the RB209 nutrient-recommendation domain, the tool surface is remarkably complete: it includes the main recommendation calculator plus lookups for crops, grass, soil, organic materials, rainfall, measurements, previous cropping, and fertiliser prices. There are no obvious dead-end workflows; even caveats around no-soil-analysis and country-specific paths are covered.
MCP server exposing supplements database used by iNutriPlan.com
MCP server for the Seline Analytics API
Public, read-only MCP server for FarmNeural company facts, packages, and capabilities.
MCP server for Codat — companies, connections, invoices, bills and financial statements.
