# DOE-MCP

**MCP servers over public US Department of Energy and national-laboratory
data. Every answer names the systems it came from, when they were read, and
what it did not cover.**

> **Not affiliated with the US Department of Energy.** DOE-MCP is an
> independent, community-run project. It is not endorsed by, funded by, or
> connected to DOE, the NNSA, any program office, any national laboratory, or
> any Power Marketing Administration. "DOE" names the data, not the origin.
> No answer this software returns is an official government statement.

DOE and its 17 national laboratories publish a large body of public data, across separate catalogs and APIs. This project exposes selected sources by **data domain** over a declarative source registry, and puts a
provenance envelope on every answer so a caller can check where it came from
and what it does not cover.

- **Use it.** [Install](#install), connect a client, ask a question. The
  literature vertical needs no credential. See [examples/README.md](examples/README.md)
  for workflows across all five servers.
- **Contribute.** The most useful contribution is a source manifest, not
  code. [CONTRIBUTING.md](CONTRIBUTING.md) has the rules for people;
  Local coding-agent rules live in `AGENTS.md`.
- **Understand it.** [The workflow guide](docs/guide.md) explains the
  runtime, coverage and evaluation. [docs/reference.md](docs/reference.md)
  lists every tool and argument, read back from the assembled servers.

## Status

Early. Five servers work against live endpoints: the research vertical
(literature, datasets, software, rulemakings, and NERSC's status
board behind its facilities toolset), the energy-data vertical (EIA, BPA,
EAGLE-I's outage releases, the wind and solar inventories, fuel economy),
the earth server (Daymet, ESS-DIVE, the ESGF index, the Sage node
inventory), the materials server (OPTIMADE structures, basis sets), and the
bio server (NMDC's microbiome studies and biosamples). The other science
domains are registry inventory with named next steps. The table below is
generated from the registry the servers run on, and CI fails when it
drifts. [The workflow guide](docs/guide.md) names the current demo and
evaluation limits.

<!-- status:begin -->
<!-- Generated by tools/build_site.py from the registry the servers run on. Edit the registry, not this block. -->
| | |
|---|---|
| Servers shipping | `doe-research` (11 tools, `research:default`), `doe-energy-data` (11 tools, `energy:default`), `doe-earth` (8 tools, `earth:default`), `doe-materials` (7 tools, `materials:default`), `doe-bio` (5 tools, `bio:default`) |
| Registry | 83 source manifests, 24 active, 54 organizations (17 national laboratories), 5 server entries and 1 skill pack in the plug-in catalog |
| Capabilities | 37 of 53 in the vocabulary served by an active source |
| Adapters | `basis_sets`, `curated`, `daymet`, `eia_v2`, `esgf`, `essdive`, `facility_status`, `federal_register`, `fueleconomy`, `json_document`, `nmdc`, `opendatasoft`, `optimade`, `osti_family`, `postgrest`, `sage`, `self_registry`, `text_feed`, `vips` |
| Verified live | Sage / Waggle edge sensor network, Basis Set Exchange, BPA transmission operations feeds, EAGLE-I historical outage datasets, DOE Software Inventory (code.json), DOE Open Data Catalog (data.json), EIA API v2, ESGF search (CMIP6), ESS-DIVE (Environmental System Science Data Infrastructure), Federal Register API (Department of Energy documents), fueleconomy.gov web service, Materials Project OPTIMADE endpoint, NERSC Superfacility API (system status), National Microbiome Data Collaborative Runtime API, ORNL DAAC (Daymet, MODIS/VIIRS subsets), Open Energy Data Hub (ORNL), DOE Data Explorer, DOE CODE, DOE PAGES (Public Access Gateway for Energy and Science), OSTI.GOV Search API, US Large-Scale Solar Photovoltaic Database (USPVDB), US Wind Turbine Database (USWTDB) |
| Not verified | none |
| Records reachable with no credential | 20,153,893 |
| Registry revision | 2026-10-20 |
<!-- status:end -->

## Install

Not on PyPI yet. Install from a checkout of this repository, from its root:

```bash
pipx install .
```

For development, with the test dependencies:

```bash
uv venv --python 3.12 && uv pip install -e ".[all]" --group dev
```

Then check the install and register the servers with a client:

```bash
doe-mcp doctor
doe-mcp configure claude-code        # or claude-desktop, vscode, cursor
```

`doe-mcp configure claude-code --all` registers every server at once. Where a
source needs a key, it goes in your own credentials file through `doe-mcp
configure credentials`, never into a client config.

## Recorded walkthroughs

[Inspect the demos](docs/demos.html): literature evidence, coverage gaps,
grid series, facility coordinates, environmental data and materials.
The geothermal dataset-card starter and microbiome sample finder include
metadata, source evidence and explicit gaps. Run the recorded workflows
locally with no key or model call:

```bash
doe-mcp demo evidence --output ./evidence-run
doe-mcp demo dataset-card --output ./dataset-draft
doe-mcp demo microbiome --output ./microbiome-samples
doe-mcp demo client
doe-mcp skills
```

[The workflow guide](docs/guide.md) covers live execution, installed skills
and optional model evaluation. Recorded examples do not establish current
publisher availability or measured agent accuracy.

## Try it from the command line

```bash
doe-mcp tools call research.search_literature \
  --args '{"query": "perovskite tandem solar cell", "year_from": 2024, "rows": 3}'
```

The answer is a complete envelope: the records under `data`, the OSTI
collections that were searched under `provenance`, one `evidence` entry per
record, and `coverage.pagination` set to `truncated`, because far more
matches exist than the three returned. No credential is involved.

## Questions it takes

| Question | Tool |
|---|---|
| What DOE-funded work exists on perovskite tandem cells? | `research.search_literature` |
| Where is the dataset behind this paper, and what is its DOI? | `research.search_datasets` |
| Has a national laboratory released Python code for lattice QCD? | `research.search_software` |
| Can I read the full text, and where? | `research.get_record` |
| What does Oak Ridge publish? | `registry.lab_crosswalk` |
| I have an old NREL URL that 404s. What happened? | `registry.resolve_org` |
| Which DOE catalogs hold geothermal datasets? | `discovery.search_all_catalogs` |
| What are DOE's own designated durable data resources? | `discovery.list_pure_resources` |
| Does DOE-MCP cover geothermal, and if not, why not? | `registry.search_sources` |
| How much electricity did CAISO demand last night? | `energy.grid_status` |
| What does this car actually get? | `fuel.find_vehicle` |
| What is BPA's load and wind generation right now? | `grid.get_bpa_operations` |
| Which EAGLE-I release holds county outage history for 2021, and where is it? | `grid.find_outage_history` |
| Which wind turbines and solar farms are inside this bounding box? | `facility.find_wind_turbines`, `facility.find_solar_facilities` |
| What has DOE proposed or finalized on appliance standards this year? | `docs.search_rulemakings` |
| Why is the VIPS patent index unavailable? | `registry.describe_source` (`pnnl-vips`) |
| What was the weather at these coordinates every day last summer? | `earth.get_daymet_point` |
| What data exists from the NGEE Arctic campaign, and who produced it? | `earth.search_datasets`, `earth.get_dataset` |
| Which climate models ran the historical experiment for surface temperature? | `climate.discover_facets`, `climate.search_cmip` |
| Where are the air-quality sensors near this city, and what is on them? | `sensors.find_nodes` |
| Is Perlmutter up, and when is NERSC's next maintenance? | `compute.facility_status` (in the `research:discovery` profile) |
| Is there soil microbiome data from Washington State, and who collected it? | `bio.search_biosamples`, `bio.get_study` |

## Architecture

DOE-MCP layers agent workflows over focused domain servers, typed tools, and
GET-only publisher adapters:

```mermaid
graph TD
    User[AI Client / User] -->|Capability Walk| Skills[Skills: Task Guides]
    User -->|MCP Protocol / stdio| Servers[Domain Servers]
    Skills -.->|Composes Capabilities| Servers

    subgraph "Shipping MCP Servers (Profiles: 8-12 tools default, <=20 ceiling)"
        S1["doe-research<br/>literature, datasets, software, rulemakings"]
        S2["doe-energy-data<br/>EIA-930 grid, BPA operations, outage history, facilities"]
        S3["doe-earth<br/>Daymet weather, ESS-DIVE, ESGF, Sage"]
        S4["doe-materials<br/>Materials Project OPTIMADE, Basis Sets"]
        S5["doe-bio<br/>NMDC microbiome studies and biosamples"]
    end

    Servers --> S1
    Servers --> S2
    Servers --> S3
    Servers --> S4
    Servers --> S5

    subgraph "Core & Adapters (Read-Only GET)"
        Registry[Declarative source registry]
        Adapters[Publisher adapters: pagination, rate limits, dialects]
    end

    S1 & S2 & S3 & S4 & S5 --> Registry
    S1 & S2 & S3 & S4 & S5 --> Adapters

    Adapters -->|HTTP GET / Dialect Translation| Upstream[Public DOE, National Lab & Federal APIs]
```

Tool counts per profile, the number of manifests and the adapter list are
in the generated status block above.

## The envelope

Every tool returns `{data, provenance, evidence, coverage, warnings, ...}`.
The part that matters most is `coverage`, which has five independent
dimensions that are never collapsed into one status:

```json
{"registry": "covered", "execution": "complete", "pagination": "truncated",
 "source_claim": "partial", "result": "hit"}
```

`result: "empty"` with `registry: "covered"` means the searched systems hold
no record. `registry: "none"` means DOE-MCP has no source for that question
and the data may well exist. Conflating those two is the most damaging thing
a government-data tool can do to someone relying on it.

DOE's additions to the inherited contract: a **citation block** (several
publishers require citation as a condition of use), a required
**`dataset_version`** (ATB 2024 and ATB 2026 are different data with the same
name), typed **access recipes** for data too big to inline, and three warning
codes for confusions this ecosystem actually produces: `derived_layer`,
`catalog_vintage`, `domain_migrated`.

## Licensing

Code Apache-2.0. **The source registry under `sources/` is CC0**: the
inventory of what exists, who stewards it, and what its terms say is more
useful unencumbered. Docs CC-BY-4.0. See [NOTICE](NOTICE).

## Lineage

The packaging shape (one pip package, many focused stdio servers, the
doctor/configure CLI) is borrowed with thanks from **PNNL's NEPA-MCP**. That
project does not endorse this one.