Skip to main content
Glama
Oceanum MCP

Oceanum MCP

Official
by oceanum-io

Oceanum MCP

An MCP server package that provides AI assistants with access to the Oceanum platform for ocean/environmental data and cloud storage.

Servers

This package contains multiple MCP servers, selectable at runtime:

Server

Description

datamesh

Search, query, and manage ocean/environmental datasets

storage

List, read, write, and delete files in Oceanum cloud storage

combined

All tools from both servers under a single endpoint (default)

Related MCP server: Code Ocean MCP Server

Prerequisites

Get an API token from oceanum.io. Set it as the DATAMESH_TOKEN environment variable (or keep it out of the client config: see Supplying the token without putting it in the client config).

Installation

Requires Python 3.10–3.13. Python 3.14 is not supported yet: the zarr 2.x stack that oceanum depends on needs numcodecs<0.16, which has no Python 3.14 wheels.

pip install oceanum-mcp

Or run directly with uvx. Pass --python 3.13 so uv doesn't pick Python 3.14 (uv downloads 3.13 if it isn't installed):

uvx --python 3.13 oceanum-mcp              # combined server (default)
uvx --python 3.13 oceanum-mcp datamesh     # datamesh only
uvx --python 3.13 oceanum-mcp storage      # storage only
uvx --python 3.13 oceanum-mcp --list       # show available servers

Configuration

Claude Desktop

Add to your claude_desktop_config.json:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

Combined server (all tools):

{
  "mcpServers": {
    "oceanum": {
      "command": "uvx",
      "args": ["--python", "3.13", "oceanum-mcp"],
      "env": {
        "DATAMESH_TOKEN": "your-token-here"
      }
    }
  }
}

Individual server (datamesh only):

{
  "mcpServers": {
    "oceanum-datamesh": {
      "command": "uvx",
      "args": ["--python", "3.13", "oceanum-mcp", "datamesh"],
      "env": {
        "DATAMESH_TOKEN": "your-token-here"
      }
    }
  }
}

Claude Code

# Combined server
claude mcp add --transport stdio oceanum -- uvx --python 3.13 oceanum-mcp

# Individual server
claude mcp add --transport stdio oceanum-datamesh -- uvx --python 3.13 oceanum-mcp datamesh

Set the token in your environment:

export DATAMESH_TOKEN=your-token-here

VS Code / Cline / Continue

Use stdio transport with the same command:

{
  "command": "uvx",
  "args": ["--python", "3.13", "oceanum-mcp"],
  "env": {
    "DATAMESH_TOKEN": "your-token-here"
  }
}

If the server reports that the token "looks like an unresolved client variable", your client passed a config variable such as ${input:datamesh-token} through literally instead of substituting it. Restart the server from the client so it prompts for the input again, or use one of the options below.

Supplying the token without putting it in the client config

For the stdio transport the server takes its token from the first of these that is set:

  1. DATAMESH_TOKEN — the token itself.

  2. DATAMESH_TOKEN_FILE — path to a file containing the token (~ is expanded). On macOS/Linux the file must be accessible only by you; the server refuses it otherwise:

    install -m 600 /dev/null ~/.config/oceanum/token
    echo "your-token-here" > ~/.config/oceanum/token
    "env": { "DATAMESH_TOKEN_FILE": "~/.config/oceanum/token" }

    The file is read on every call, so a replaced token takes effect without restarting the server.

  3. DATAMESH_TOKEN_KEYRING=1 — read the token from the OS credential store (macOS Keychain, Secret Service/libsecret on Linux, Windows Credential Manager) via the keyring package, under service oceanum-mcp, user datamesh-token. Store it once with:

    uvx keyring set oceanum-mcp datamesh-token

    The token is looked up once per server process, so restart the server after changing it. keyring currently arrives as an indirect dependency of fastmcp, which is not guaranteed to last: to be sure, install the oceanum-mcp[keyring] extra (with uvx: uvx --python 3.13 --from "oceanum-mcp[keyring]" oceanum-mcp).

Troubleshooting: -32022 on the first initialize

A client may log "connection is serving the 2026-07-28 protocol; the initialize handshake is not accepted" (error -32022) once on connect. This comes from the MCP Python SDK, not from this server's code: in SDK 2.3.0 (mcp/server/runner.py, serve_dual_era_loop) a connection's first request fixes its protocol era. A client that opens with a 2026-07-28 request and then sends the older initialize handshake on the same connection is refused, with the supported versions in the error. Clients are expected to retry with a supported version; GitHub Copilot did so and connected.

Environment Variables

Variable

Required

Description

DATAMESH_TOKEN

Yes*

Oceanum API token (shared by all servers)

DATAMESH_TOKEN_FILE

Yes*

Path to a file holding the token, readable only by its owner; used when DATAMESH_TOKEN is unset

DATAMESH_TOKEN_KEYRING

Yes*

1/true/yes/on to read the token from the OS credential store (once per process); used when neither of the above is set

DATAMESH_SERVICE

No

Custom datamesh service URL (default: https://datamesh.oceanum.io)

STORAGE_SERVICE

No

Custom storage service URL (default: https://storage.oceanum.io)

OCEANUM_DOMAIN

No

Override the base domain for all services (default: oceanum.io)

OCEANUM_ADMIN_SERVICE

No

Admin service URL, which serves datasource notices; must be an https:// origin (or http://localhost/127.0.0.1) with no user info, path, query or fragment, else notices are off. Unset, it is https://admin.oceanum.io or https://admin.oceanum.tech for those Datamesh services, and there are no notices with any other Datamesh service

OCEANUM_MCP_DISABLE_NOTICES

No

1/true/yes/on to stop get_datasource_info looking up datasource notices (get_datasource_notices still works)

OCEANUM_MCP_READ_ONLY

No

1/true/yes/on to disable write tools (update_metadata, storage write_file/delete_file); any value other than these or 0/false/no/off stops the server at startup

OCEANUM_MCP_MAX_INLINE_BYTES

No

Max staged result size returned inline by query_data (default 50,000,000)

OCEANUM_MCP_MAX_INLINE_ROWS

No

Max rows/records previewed inline before truncation (default 100)

OCEANUM_MCP_EXPORT_DIR

No

If set, export_query may only write inside this directory

OCEANUM_MCP_STAGE_TIMEOUT

No

Seconds a Datamesh staging/query request may wait on the gateway before the tool returns a timeout error (default 120)

OCEANUM_MCP_AUTH

No

Auth scheme for --transport http: auto (default), datamesh, auth0, or none

OCEANUM_MCP_AUTH0_DOMAIN

No

Auth0 tenant domain for auth0 mode (default: auth.oceanum.io)

OCEANUM_MCP_AUTH0_AUDIENCE

No

Auth0 API audience for auth0 mode (default: https://api.oceanum.io)

OCEANUM_MCP_PUBLIC_URL

No

Externally visible base URL (e.g. https://mcp.oceanum.io); enables OAuth discovery metadata for claude.ai connectors

OCEANUM_MCP_CORS_ORIGINS

No

Comma-separated browser origins allowed by CORS on http (default: https://*.oceanum.io,https://*.oceanum.tech,https://vscode.dev; empty string disables)

* One token source (DATAMESH_TOKEN, DATAMESH_TOKEN_FILE or DATAMESH_TOKEN_KEYRING) is required for the stdio transport (and for --transport http with OCEANUM_MCP_AUTH=none); in authenticated http mode each request carries its own credential and no server-side token is needed.

Hosted mode (HTTP transport)

Run any server as a shared, multi-tenant HTTP service:

oceanum-mcp datamesh --transport http --host 0.0.0.0 --port 8000

This serves the MCP streamable-HTTP endpoint at http://<host>:<port>/<server> (/datamesh here; override with --path). Each server owning its own path lets several MCP servers share one domain behind an ingress — e.g. https://mcp.oceanum.io/datamesh and https://mcp.oceanum.io/storage. Every request must present a bearer credential, and all Datamesh/Storage calls are made as that request's user — connections are cached per credential and never shared between tokens.

Auth schemes (OCEANUM_MCP_AUTH):

  • auto (default) — accepts either credential, detected per request: Datamesh tokens arrive in the X-DATAMESH-TOKEN header (the same header the gateway itself uses), Auth0 JWTs in Authorization: Bearer. A Datamesh token sent as the bearer also works (JWT-shape detection routes it), but the dedicated header is the canonical form. Each request runs as the identity its own credential resolves to.

  • datamesh — only Datamesh tokens are accepted. The server validates them against the gateway. Add to Claude Code with:

    claude mcp add --transport http oceanum-datamesh https://mcp.oceanum.io/datamesh \
      --header "X-DATAMESH-TOKEN: <datamesh-token>"
  • auth0 — clients send an Auth0-issued JWT, validated against the tenant JWKS and forwarded to the Datamesh gateway as-is.

  • none — no authentication; every request uses the server's own DATAMESH_TOKEN. Only for trusted-network deployments.

Notes:

  • export_query works over HTTP but changes shape: instead of writing to the server's local filesystem (meaningless for remote clients), it returns a time-limited, self-authenticating gateway download_url that the caller fetches out-of-band. The path argument is ignored on hosted servers. The URL needs no credential, so it is a capability — treat it as a secret. Gateways with scoped download links (OCE-317) also limit its lifetime and number of fetches; the response then reports them as expires_at and max_uses. Hosted exports are capped at 10 GB to bound what one link can pull, and exports above 1 GB need confirm_size_bytes (see export_query).

  • A rejected Datamesh token (in X-DATAMESH-TOKEN, or a non-JWT Authorization: Bearer) gets a 401 saying the token was not accepted, without the OAuth resource_metadata challenge (which would send clients such as VS Code into a sign-in flow that cannot fix a Datamesh token). Requests with no credential, and failed JWT (OAuth) bearers, still get the challenge, so OAuth discovery and re-authorization are unaffected. If the token could not be verified because the Datamesh service was unreachable, the response is a 503 asking to retry instead. A credential that is an unsubstituted client variable (${input:...}) gets a 401 saying so (except with OCEANUM_MCP_AUTH=none, where client credentials are ignored).

  • Combine with OCEANUM_MCP_READ_ONLY=1 to run a read-only public service.

  • Pass --stateless when running behind a load balancer or on autoscaled platforms (Cloud Run, etc.): sessions are otherwise held in instance memory, and consecutive requests routed to different instances would fail.

  • Breaking change for --transport sse (deprecated): sse is a network transport and now behaves like http — authenticated by default and no export_query. Set OCEANUM_MCP_AUTH=none to restore the old unauthenticated behavior.

Serving under an external ASGI server

To run under uvicorn/gunicorn (multiple workers, serverless platforms), use the packaged app factory — it applies the same auth and tool policy as the CLI; serving mcp.http_app() directly would bypass both:

uvicorn --factory oceanum_mcp.app:create_http_app --host 0.0.0.0 --port 8000

Requests on a network transport never fall back to the server's DATAMESH_TOKEN: an unauthenticated request fails unless OCEANUM_MCP_AUTH=none was set explicitly.

Browser clients (CORS)

Browser-based MCP clients send a CORS preflight (OPTIONS) before every cross-origin request. The http app answers preflights for an allowlist of origins, ahead of authentication (a preflight never carries a credential); all other requests still require one. Configure the allowlist with OCEANUM_MCP_CORS_ORIGINS, comma-separated:

OCEANUM_MCP_CORS_ORIGINS="https://*.oceanum.io,https://vscode.dev,http://localhost:6274"
  • Unset: https://*.oceanum.io, https://*.oceanum.tech and https://vscode.dev.

  • Empty string: CORS disabled — no CORS headers, preflights are not answered.

  • Each entry is an exact origin (scheme://host[:port], no path). A leading *. matches exactly one subdomain label: https://*.oceanum.io allows https://app.oceanum.io but not https://oceanum.io, https://a.b.oceanum.io or https://oceanum.io.evil.com. Bare *, single-label wildcards (https://*.io) and malformed entries are rejected at startup; a scheme-default port (:443 on https) is dropped, as browsers omit it. Multi-label public suffixes (https://*.co.uk, https://*.github.io) cannot be detected — never configure them: they would admit every site under that suffix.

  • Credentials travel in Authorization / X-DATAMESH-TOKEN headers, never cookies, so Access-Control-Allow-Credentials is never sent. Mcp-Session-Id and WWW-Authenticate are exposed to browser scripts.

claude.ai custom connectors

Set OCEANUM_MCP_PUBLIC_URL to the server's public base URL and the auto and auth0 modes additionally serve OAuth 2.0 Protected Resource Metadata (RFC 9728) naming the Auth0 tenant, plus a WWW-Authenticate challenge on 401s — which is how claude.ai discovers where to send users to authorize. The Auth0 tenant must allow client registration for Claude (Dynamic Client Registration, or a pre-registered application) and issue JWT access tokens for the configured audience. Header/bearer clients are unaffected either way.

Post-deploy smoke test

scripts/smoke_hosted.py opens real gridded datasources over zarr through a deployed server (load_datasource, and a whole-grid query_data), which catches a broken data stack in the image that tools/list does not. There is no deploy pipeline (the hosted service is deployed by hand with the local deploy_cloudrun*.sh scripts), so run it manually after each deploy:

DATAMESH_TOKEN=<token> uv run python scripts/smoke_hosted.py https://mcp.oceanum.io/datamesh

The token must be valid for the gateway the server uses. --load-datasource, --query-datasource and --query-variable override the default datasets. The URL must be https unless the host is local, since the token is sent with every request. Exit status: 0 all checks passed; 1 a check failed, or the server was unreachable or rejected the token; 2 usage error (DATAMESH_TOKEN unset, or a non-https remote URL); 3 inconclusive (a datasource was not found, not shared or empty, the query was refused, or it was small enough to be downloaded instead of opened over zarr).

The hosted image installs this package from the source checkout (pip install .), not from PyPI, so dependency fixes take effect on the next deploy from main without a release.

Datamesh Tools

The intended workflow is: search_catalog → get_datasource_info → stage_query (dry run: learn the result size without downloading) → query_data for small results inline, or export_query to write large results to a file that analysis code reads directly.

search_catalog

Search the Datamesh catalog with optional text search, time range, and bounding box filters. Returns a JSON object with count and results; if count equals limit, more matches may exist. Only datasources accessible to the request's credentials are searched. An empty or short (below limit, with nothing omitted) result carries a fixed scope note saying a missing datasource either does not exist or is not accessible with these credentials; it never reveals whether, how many, or which inaccessible datasources exist.

By default each result is a compact summary: id, name, a description cut to 200 characters, tstart/tend, bounds (rounded outward to 4 decimals), variable names when the catalog provides them (first 20, with variables_total), and the first 5 tags (each clipped to 60 characters, with tags_total). Unusable bounds (empty or non-finite) are null. detail="full" returns each datasource's complete catalog record instead. The response is compact JSON, and its total size is bounded (~20k characters for summaries, ~100k for full records): matches beyond the bound are dropped, counted in omitted, and a note asks to refine the search.

The catalog ranks datasources by fuzzy similarity to the search text and returns the closest ones even when none contains it. With search text, the response therefore carries text_matches: how many results have every search word (as a whole word, singular or plural) in their id, name, description, tags, labels or variable names. or separates alternatives and negated words or phrases (-era5, -"era5 grid", not era5) are ignored. This count is only a signal: no result is dropped for failing it, but results that fail are marked "text_match": false, and a text_note ahead of the results explains. If no result contains the text as whole words (text_matches: 0), every result is still returned, but in summary mode as id and name only, with a text_note first and no scope note.

Parameter

Type

Description

search

string

Text search for name, description, or tags

time_start

string

ISO 8601 start time

time_end

string

ISO 8601 end time

bbox

list[float]

Bounding box [xmin, ymin, xmax, ymax] in WGS84

limit

int

Max results to return (default 20)

detail

string

summary (default) or full

get_datasource_info

Get the metadata for a datasource: coverage, schema dims and coordinates, variables, and attributes.

The default (detail="summary") is a compact view with the same fields as the full record. Each variable and coordinate keeps its dims, shape, dtype and units/long_name/standard_name attributes only (other attributes are counted in attrs_omitted); the description is clipped to 1,000 characters, global attributes and info are capped at 25 entries (with *_omitted counts), and long values are clipped to 200 characters. schema.data_vars/schema.attrs are omitted as duplicates of variables/attributes. Past ~30k characters the summary sheds, in order and only as far as needed: global attributes, coordinate attributes, variable attributes other than units, variable shape/dtype, and finally per-variable details; the note names what was shed. Every variable name is always listed. detail="full" returns the complete record. Both are compact JSON.

An unknown id adds suggestions: visible datasources whose id or name differs from the requested one by a typo (at most one character edit per 7 characters, up to 3, with the differing words still similar), or extends a request of two or more words. An id that merely shares generic words such as oceanum_wave_..._era5, or a single word such as wave, is never suggested.

Active notices about the datasource (see get_datasource_notices) come first in both views: notices lists the outage and action_required ones (e.g. a deprecation), outages first, then newest first, at most 5, as severity, title, body (markdown links and images reduced to their text, clipped to 500 characters) and expires_at, under a fixed notices_note that labels them as announcements rather than instructions; notices_hint counts any others ("at least N more" when the service sent its maximum of 20). These keys are absent when there are none, and the record cannot supply them: a record field with one of their names is renamed record_<name>. The summary view's size budget includes them.

The lookup never fails or holds up the tool: it runs alongside the metadata request and is given at most 3 s in all; errors count as "no notices". There is no lookup for an empty id. Set OCEANUM_MCP_DISABLE_NOTICES=1 to turn it off.

Answers are cached per process, per credential and datasource, and shared with get_datasource_notices; concurrent lookups for the same pair share one request. A list of notices is reused for 10 minutes; no notices (an empty list, or a 404: none visible) for 2 minutes, so a newly published notice shows soon. A failure holds off further lookups of that datasource for 1 minute (or the service's Retry-After, up to 10 minutes); a failure of the service itself (a timeout, a connection error, a 5xx or 429) holds off requests to it for every datasource, so a hung service costs one timeout, not one per datasource. A refused credential or a 404 does not. Meanwhile the last good answer is used, however old; with none, get_datasource_info shows no notices. A failure never replaces the last good answer, any success ends the hold-off, and running out of time before the request is sent is not remembered. Notices that have expired, or are not yet published, are dropped from every answer, cached or not. In the summary, a notice whose title is only a link or image is left out (and counted in notices_hint).

Parameter

Type

Description

datasource_id

string

Datasource ID

detail

string

summary (default) or full

get_datasource_notices

Get the active notices about a datasource — deprecations, outages, changes and other announcements published by Oceanum — in full, most severe first (outage, action_required, then the rest), newest first within each, at most 20. Returns {"datasource_id", "notices_note", "notices": [{severity, title, body, publish_at, expires_at}]}; severity is one of info, change, action_required or outage, and body is markdown. notices is empty when there are none, or when the datasource does not exist or is not visible with your credentials. Any failure (no known admin service, no credential, the service unreachable, slow, busy or rejecting the credential) returns JSON with an error, plus retry_after_s when the service asked to back off. While a recent failure holds off lookups (see get_datasource_info), the tool answers from the last good answer without asking the service, or, with none, with an error and retry_after_s: it never reports "no notices" for a lookup that did not happen.

Notices come from the admin service (OCEANUM_ADMIN_SERVICE), called with the same credential as the other tools: a Datamesh token as X-DATAMESH-TOKEN, or in hosted mode an Auth0 JWT as Authorization: Bearer. So the credential only goes to a known admin service: unless OCEANUM_ADMIN_SERVICE is set, notices are looked up only for the datamesh.oceanum.io and datamesh.oceanum.tech services, and with any other Datamesh service (local, staging) no request is made. Redirects are never followed; a redirect or refused credential is logged as a warning (at most every 10 minutes). Notice text is treated as untrusted whatever the service sends: malformed notices are dropped, control and formatting characters stripped, titles clipped to 200 characters and bodies to 2,000. The lookup is given at most 3 s.

The notice text is still text an Oceanum administrator wrote, and it reaches the model: the fixed notices_note labels it as announcements to pass on, not instructions, and the summary in get_datasource_info drops links, but nothing can stop a model from following instruction-like text it reads. Notices are only as trustworthy as the admin service and the people who publish them; OCEANUM_MCP_DISABLE_NOTICES=1 keeps them out of get_datasource_info.

Parameter

Type

Description

datasource_id

string

Datasource ID

stage_query

Dry-run a query on the Datamesh gateway: reports the result size, container type, and domain length without downloading any data, echoes the canonical query, and recommends the next step (inline query vs export vs narrowing). The inline recommendation counts records as well as bytes: a result with more records than OCEANUM_MCP_MAX_INLINE_ROWS would come back from query_data as a preview, so it is steered to resampling, aggregation or export_query instead. Accepts the same query parameters as query_data.

query_data

Query a datasource with filters and return small results inline as coordinate-attributed JSON records with explicit truncated/lazy flags. The query is staged first: gridded results above the inline limit are summarized lazily (structure only); tabular results above the limit, and any result above 1 GB, are refused with the staged size and alternatives (the 1 GB refusal also lists each variable's dims). The exception is limit combined with time_resolution or aggregate_operations: the limited slice may be small, so it is still opened lazily. Library warnings (e.g. the 2,000,000-row cap on tabular queries) are included in the response.

Parameter

Type

Description

datasource_id

string

Datasource to query

variables

list[string]

Variables to select

time_start

string

ISO 8601 start of a time range (open-ended if omitted)

time_end

string

ISO 8601 end of a time range (open-ended if omitted)

times

list[string]

Discrete times (series selection); excludes time_start/time_end

time_resolution

string

Server-side temporal downsampling (pandas frequency, e.g. 1h, 1D, 30D; not MS/ME/QS/YS or finer than 1 minute)

time_resample

string

Resampling method for time_resolution: mean, nearest, linear

bbox

list[float]

Bounding box [xmin, ymin, xmax, ymax]

geofilter_feature

object

GeoJSON Feature (Point, MultiPoint, or Polygon) for selection

geofilter_interp

string

Interpolation for feature selection: nearest or linear

geofilter_resolution

float

Max spatial resolution for downsampling, in CRS units

level_min

float

Minimum vertical level

level_max

float

Maximum vertical level

levels

list[float]

Discrete vertical levels (series selection)

level_interp

string

Interpolation for level series: nearest or linear

coord_filters

list[object]

Coordinate selections: [{"coord": "name", "values": [...]}]

crs

string/int

CRS for filter coordinates and returned data

aggregate_operations

list[string]

Aggregation ops: mean, min, max, std, sum

aggregate_spatial

bool

Aggregate over spatial dims (default true)

aggregate_temporal

bool

Aggregate over temporal dims (default true)

limit

int

Keep the last N records (see below)

limit keeps the last N records (Datamesh semantics: the last N steps along time/ensemble), with the same meaning in stage_query, query_data, and export_query. Combined with time_resolution or aggregate_operations it is applied by this server after resampling/aggregation instead of being sent to Datamesh (which would apply it to the native records first), so stage_query reports the unlimited size as an upper bound. Hosted export_query cannot apply it to a gateway download and refuses that combination: drop limit or narrow the time range.

export_query

Export the full result of a query — the data-handle path for results too large to return inline. Behaviour depends on transport:

  • Local (stdio): writes the result to the local file path (required) and returns that path. Gridded datasets stream lazily to NetCDF; tabular results write Parquet or CSV.

  • Hosted (http/sse): returns a time-limited, self-authenticating gateway download_url (choose format; path is ignored). Fetch it out-of-band. When the response includes expires_at or max_uses (gateways with scoped links), fetch it before expires_at and at most max_uses times. Results above 10 GB (for CSV, the estimated text size, about 3x the staged size) are refused (narrow the query); above 2 GB the link is returned with a large-download warning.

export_query sizes the result itself, whether or not stage_query was called, and refuses before anything is exported. Results above the export cap are refused outright: the cap limits the transfer, so splitting the request into parts is not a way round it. An export that writes more than 1 GB in the chosen format (CSV text is estimated at about 3x the staged size; locally, a dataset written as CSV/Parquet is sized as the flattened table, and a limited one on its limited slice) is refused unless confirm_size_bytes equals that size. On hosted servers a value within 5% is accepted, because the download stage is sized separately from stage_query's stage. The refusal reports the exact value as export_size_bytes. The agent should pass it only after the user has agreed. The server cannot see whether the user was asked, so the check is advisory: it forces an explicit, size-aware step for each large export.

Accepts the same query parameters as query_data plus:

Parameter

Type

Description

path

string

Destination file path (parent directories are created)

format

string

netcdf (datasets), parquet or csv (tabular); sensible default

overwrite

bool

Overwrite an existing file (default false)

confirm_size_bytes

int

Bytes (in the chosen format) the user agreed to export; required above 1 GB; exact match (within 5% on hosted)

load_datasource

Summarize an entire datasource. Gridded datasources are opened lazily (no data download); tabular datasources are downloaded only if under the inline size limit.

Parameter

Type

Description

datasource_id

string

Datasource to load

update_metadata

Update metadata on an existing datasource. Only provided fields are changed. Disabled when the server runs with OCEANUM_MCP_READ_ONLY set.

Parameter

Type

Description

datasource_id

string

Datasource to update

name

string

New name

description

string

New description

tags

list[string]

New tags

labels

list[string]

New labels

info

object

Additional metadata object

details

string

URL for datasource details

Storage Tools

list_files

List files and directories in Oceanum cloud storage.

Parameter

Type

Description

path

string

Directory path to list (default: "/")

recursive

bool

List subdirectories recursively

file_exists

Check if a file or directory exists in storage.

Parameter

Type

Description

path

string

Path to check

read_file

Read the contents of a text file from storage.

Parameter

Type

Description

path

string

Path to the file

write_file

Write text content to a file in storage.

Parameter

Type

Description

path

string

Destination path

content

string

Text content to write

delete_file

Delete a file or directory from storage.

Parameter

Type

Description

path

string

Path to delete

recursive

bool

Delete directory contents recursively

file_info

Get metadata about a file or directory.

Parameter

Type

Description

path

string

Path to inspect

Example Workflows

Discover wave data in the Pacific:

  1. search_catalog(search="wave", bbox=[120, -50, 180, 10])

  2. get_datasource_info(datasource_id="some-wave-dataset")

  3. stage_query(datasource_id="some-wave-dataset", variables=["Hs", "Tp"], time_start="2024-01-01", time_end="2024-01-31") to check the result size

  4. query_data(...) with the same parameters if small, or export_query(..., path="waves.nc") if large

Shrink a 40-year hourly time series to something inline-sized:

  1. stage_query(datasource_id="hindcast", variables=["Hs"], time_start="1984-01-01", time_end="2024-01-01") — too large

  2. query_data(..., time_resolution="30D", time_resample="mean") — roughly monthly means, small enough to return inline (calendar aliases such as "1MS" are refused until a Datamesh engine fix ships: the engine currently reads "1MS" as 1 millisecond)

Browse and read files in cloud storage:

  1. list_files(path="/") to see top-level contents

  2. list_files(path="/my-project", recursive=True) to drill down

  3. read_file(path="/my-project/config.json") to read a file

Get a quick summary of a dataset:

  1. get_datasource_info(datasource_id="my-dataset") to see variables and time range

  2. query_data(datasource_id="my-dataset", limit=10) to preview the data

Available Tools

13 tools
datamesh_export_queryDatamesh Export QueryA
Idempotent

Export the FULL result of a query — the data never enters the conversation.

Behavior depends on how the server is running:

  • Hosted (http/sse): returns a time-limited, self-authenticating gateway download_url (choose format via format); path is ignored. Fetch the URL out-of-band — it needs no credential, so treat it as a secret. When the response includes expires_at or max_uses, fetch it before expires_at and at most max_uses times. Refused above the hosted export cap, and for limit combined with time_resolution/aggregate_operations.

  • Local (stdio): writes the result to the local file path (required) and returns that path. Gridded datasets stream to NetCDF; tabular results write Parquet or CSV, as do datasets small enough to flatten into a table (e.g. point time series). Refused up front if the result exceeds the local export cap or the destination's free disk space. With limit plus time_resolution/aggregate_operations, the last N resampled steps are written.

The size is checked before anything is exported, whether or not stage_query was called. An export writing more than 1 GB in the chosen format (CSV text is about 3x; locally, a dataset written as CSV/Parquet is sized as the flattened table, and a limited one on its limited slice) is refused unless confirm_size_bytes equals that size (on hosted, within 5%, since the download stage is sized separately): ask the user first. The refusal states the exact value. The server cannot see whether the user was asked, so this is advisory: it forces an explicit, size-aware step. The export caps limit the transfer, not the size of each file: never split a refused request into parts to get under them.

A variable given as the exact CF standard_name of exactly one of the datasource's variables (e.g. "sea_surface_wave_significant_height" for "hs") is resolved to that variable; the response then includes resolved_variables ({requested: used}) and echoes the resolved names.

Inputs: datasource_id (required), crs, bbox, path, +21 more. Returns: result

datamesh_get_datasource_infoDatamesh Get Datasource InfoA
Read-only

Get the metadata for one datasource: coverage, coordinates and variables.

The default view keeps every field of the full record but reduces each variable and coordinate to its dims, shape, dtype, units, long_name and standard_name, clips the description and long attribute values, caps global attributes. On very large records it sheds attributes and then per-variable details (the note says which), but every variable name is always listed.

Inputs: datasource_id (required), detail. Returns: result

datamesh_load_datasourceDatamesh Load DatasourceA
Read-only

Summarize an entire datasource.

Gridded datasources are opened lazily (no data download). Tabular datasources are downloaded only if under the inline size limit; use query_data with filters to retrieve larger ones.

Inputs: datasource_id (required). Returns: result

datamesh_query_dataDatamesh Query DataA
Read-only

Query a datasource and return small results inline.

The query is staged first; results larger than the inline limit are not downloaded (datasets are summarized lazily, tabular queries are refused with alternatives), and results above 1 GB are refused with narrowing advice and each variable's dims, unless limit is combined with time_resolution/aggregate_operations (the limited slice may be small, so a dataset is then opened lazily). Use stage_query to size a query before calling this.

Inline values are a preview when preview is true (returned of total records): never compute statistics from a preview; use aggregate_operations or time_resolution instead.

limit keeps the last N records (Datamesh semantics: the last N steps along time/ensemble). Combined with time_resolution or aggregate_operations, limit is applied by this server AFTER resampling/aggregation (to the last N resampled steps, or the last N rows of a table) rather than sent to Datamesh. A limited result is a subset of the requested range, so it is flagged as a preview, with a limit_note.

A variable given as the exact CF standard_name of exactly one of the datasource's variables (e.g. "sea_surface_wave_significant_height" for "hs") is resolved to that variable; the response then includes resolved_variables ({requested: used}) and echoes the resolved names.

Inputs: datasource_id (required), crs, bbox, limit, +17 more. Returns: result

datamesh_search_catalogDatamesh Search CatalogA
Read-only

Search the Oceanum Datamesh catalog for datasets.

Catalog search only covers datasources accessible to the credentials used for this request; a datasource missing from the results either does not exist or is not accessible with these credentials, so do not report it as nonexistent.

Inputs: bbox, limit, detail, search, +2 more. Returns: result

datamesh_stage_queryDatamesh Stage QueryA
Read-only

Dry-run a query: report the result size WITHOUT downloading any data.

Always stage before retrieving data you have not sized. The response says whether the result is small enough for query_data to return inline, and echoes the canonical query. With limit plus time_resolution or aggregate_operations the reported size is that of the unlimited result (limit is applied after resampling), so it is an upper bound.

A variable given as the exact CF standard_name of exactly one of the datasource's variables (e.g. "sea_surface_wave_significant_height" for "hs") is resolved to that variable; the response then includes resolved_variables ({requested: used}) and echoes the resolved names.

Inputs: datasource_id (required), crs, bbox, limit, +17 more. Returns: result

datamesh_update_metadataDatamesh Update MetadataA
DestructiveIdempotent

Update metadata properties on an existing datasource.

Only the provided fields will be updated; others remain unchanged. Not available when the server runs with OCEANUM_MCP_READ_ONLY set.

Inputs: datasource_id (required), info, name, tags, +3 more. Returns: result

storage_delete_fileStorage Delete FileC

Delete a file or directory from Oceanum storage.

Inputs: path (required), recursive. Returns: result

storage_file_existsStorage File ExistsB

Check if a file or directory exists in Oceanum storage.

Inputs: path (required). Returns: result

storage_file_infoStorage File InfoB

Get metadata about a file or directory in Oceanum storage.

Inputs: path (required). Returns: result

storage_list_filesStorage List FilesC

List files and directories in Oceanum cloud storage.

Inputs: path, recursive. Returns: result

storage_read_fileStorage Read FileB

Read the contents of a text file from Oceanum storage.

Inputs: path (required). Returns: result

storage_write_fileStorage Write FileA

Write text content to a file in Oceanum storage.

Creates the file if it doesn't exist, overwrites if it does.

Inputs: path (required), content (required). Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.7/5.0

    Scored across 13 tools

    Disambiguation4/5

    The data-access tools are clearly tiered by role: search_catalog finds datasources, get_datasource_info/stage_query inspect metadata and size, query_data returns inline previews, and export_query offloads large results. The only real friction is load_datasource ('summarize an entire datasource') versus get_datasource_info ('metadata for one datasource'), which can look interchangeable at a glance, plus the storage_file_exists/storage_file_info pair.

    Naming Consistency4/5

    Every name is snake_case with a clear domain prefix (datamesh_/storage_) and most follow verb_noun (query_data, export_query, stage_query, list_files, read_file, delete_file). The deviations are minor: load_datasource rather than get_datasource, and the noun-first storage_file_exists/storage_file_info.

    Tool Count5/5

    Thirteen tools split cleanly into two domains (7 datamesh, 6 storage), each earning its place in the query/export workflow. Nothing feels redundant or padded at this size.

    Completeness4/5

    The data pipeline is covered end-to-end (discover → inspect → stage → query → export → update metadata) and storage has full create/read/list/delete/stat coverage. Gaps remain: no datasource creation or deletion, no storage copy/move, and metadata updates are the only write path for the catalog.

    Maintenance

    ActivityActive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      D
      maintenance
      Access oceanographic and environmental data from 63+ ERDDAP servers worldwide through natural language queries. Search datasets, retrieve metadata, and download scientific data for climate research, marine biology, and coastal management.
      18
      -
    • A
      license
      B
      quality
      B
      maintenance
      Provides tools to search and execute Code Ocean capsules and pipelines while managing platform data assets. It enables users to interact with Code Ocean's computational resources and scientific workflows directly through natural language interfaces.
      26
      298 PyPI
      5
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI agents to interact with the OSDU® data platform through Model Context Protocol, providing tools for search, storage, schema management, and dataset operations.
      2
      Apache 2.0