generated: '2026-08-09'
method: searched
status: published
source: https://atmospore.com/article/mcp
description: Atmospore ships BOTH a hosted remote MCP server and an open-source local stdio server. The
  remote endpoint at https://mcp.atmospore.com/mcp is live and answers JSON-RPC, gated by the same Atmospore
  API key used for REST. The local server is MIT-licensed Python published from the atmospore GitHub organization.
server:
  name: atmospore
  transport: http
  url: https://mcp.atmospore.com/mcp
  protocol: JSON-RPC 2.0 (Streamable HTTP)
  authentication:
    type: apiKey
    schemes:
    - style: header
      header: Authorization
      format: Bearer ak_...
    - style: query
      parameter: key
      format: ak_...
    anonymous_access: false
    oauth: false
    note: The server error message documents the `ak_` key prefix. The setup snippet published at https://atmospore.com/article/mcp
      shows `Bearer atmo_your_key_here` instead — the docs and the server disagree on the key prefix.
      Keys come from https://atmospore.com/account.
  client_setup:
    documented_via: 'npx -y mcp-remote https://mcp.atmospore.com/mcp --header "Authorization: Bearer <key>"'
    clients_named:
    - Claude Desktop
    - Cursor
    - Continue
    - ChatGPT Desktop
  x-evidence:
    fetched: '2026-08-09'
    url: https://mcp.atmospore.com/mcp
    method: POST
    body: '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
    http_status: 401
    response: '{"jsonrpc":"2.0","error":{"code":-32001,"message":"Missing API key. Send `Authorization:
      Bearer ak_...` or `?key=ak_...`. Get a free key at https://atmospore.com/account."},"id":null}'
    interpretation: A well-formed JSON-RPC error, not an HTML shell — the endpoint is a real live MCP
      server. tools/list and initialize are both auth-gated, so live inputSchema introspection requires
      an authenticated key.
local_server:
  name: atmospore-mcp
  transport: stdio
  repository: https://github.com/atmospore/atmospore-mcp
  license: MIT
  language: python
  framework: mcp.server.fastmcp (FastMCP)
  entrypoint: atmospore-mcp
  env:
  - ATMOSPORE_API_KEY
  container: Dockerfile present in repo
  directory_listing: glama.json present (listed on the third-party Glama MCP directory)
  published_to_pypi: false
tools_source: Tool names, parameters, defaults and descriptions below are read verbatim from the first-party
  open-source server at https://raw.githubusercontent.com/atmospore/atmospore-mcp/main/atmospore_mcp/server.py
  and its README. The hosted server's live tools/list is auth-gated and was NOT introspected, so the hosted
  tool set is asserted to match the open-source server but is unverified.
tools:
- name: get_pollen
  description: Get the daily pollen forecast for a specific point on Earth. Returns a list of daily forecasts;
    each day includes date, overall_risk (Low|Moderate|High|Very High) and per-species values in grains/m3
    with individual risk levels.
  parameters:
  - name: lat
    type: number
    required: true
  - name: lon
    type: number
    required: true
  - name: forecast_days
    type: integer
    required: false
    default: 1
  source_operation: openapi/atmospore-pollen-forecasts-openapi-original.json#getPollenForecast
- name: get_top_species
  description: Get the top contributing pollen species at a point, ranked highest first. Each entry has
    species slug, display_name, max_value in grains/m3, risk_level and category.
  parameters:
  - name: lat
    type: number
    required: true
  - name: lon
    type: number
    required: true
  - name: limit
    type: integer
    required: false
    default: 5
  source_operation: openapi/atmospore-pollen-forecasts-openapi-original.json#getPollenTop
- name: get_area_average
  description: Get tree, grass and weed pollen aggregates over a radius around a point. One entry per
    day with overall_risk and per-category aggregates (tree_tot, grass_tot, weed_tot).
  parameters:
  - name: lat
    type: number
    required: true
  - name: lon
    type: number
    required: true
  - name: radius_km
    type: number
    required: false
    default: 25
  - name: forecast_days
    type: integer
    required: false
    default: 7
  source_operation: openapi/atmospore-pollen-forecasts-openapi-original.json#getPollenArea
- name: list_supported_species
  description: List all pollen species the model tracks with slug, display name, category (tree/grass/weed),
    localised names (en, no, sv) and risk-level concentration thresholds.
  parameters: []
  source_operation: openapi/atmospore-pollen-forecasts-openapi-original.json#getSpecies
resources:
- uri: atmospore://help
  description: Usage notes browsable from Claude Desktop (tool table, coordinate examples, quota notes).
error_contract:
  envelope: '{ok: false, error: <slug>, message: <string>, ...}'
  errors:
  - error: authentication_failed
    cause: bad/missing/revoked key
    hint: Check ATMOSPORE_API_KEY
  - error: quota_exceeded
    cause: daily quota hit
    fields:
    - limit
    - used
    - resets_at
    hint: Upgrade at https://atmospore.com/plans
  - error: api_error
    cause: other 4xx/5xx
    fields:
    - status
  - error: unexpected_error
    cause: unhandled exception
quotas:
  note: Every MCP tool call consumes one Atmospore API request against the key's plan quota.
  free_tier: 100 calls/day
  paid_tier: 5000 calls/day (Professional)
gaps:
- No /.well-known/oauth-protected-resource on the MCP host, so auth is undiscoverable before the first
  call.
- Docs and server disagree on the API key prefix (atmo_ vs ak_).
- The local server is not on PyPI, so its documented `pip install atmospore-mcp` fails.
deployment:
  mode: remote
  endpoint: https://mcp.atmospore.com/mcp
  verified: probed
  probe: gated
  checked: '2026-08-12'
  source: catalog MCP census