atomic-context MCP server
Commit Nov 13, 2026Provides an AI client with deterministic, citable values for physical constants, particle properties, and material properties.
Also known as askeita/atomic-context
- Stars
- 0
- Tools
- 18
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Nov 13, 2026
- License
- MIT
What is atomic-context MCP?
atomic-context provides an AI client with deterministic, citable values for physical constants, particle properties, and material properties. It exposes 18 tools, such as get_constant, search_constants, list_constant_categories.
Summary generated from the project's README
About atomic-context MCP server
Providing precise, citable physical data to AI clients is what atomic-context enables for researchers and engineers. It acts as a grounding layer for LLMs that might otherwise struggle with high-precision values, such as 12-decimal-place CODATA constants or specific heat capacities for various alloys. By connecting to an MCP-compatible client like Claude Desktop, Cursor, or Continue.dev, users can access deterministic information from verified open-source datasets.
The server runs locally using Java 21 and SQLite, supporting both stdio for desktop applications and SSE/HTTP for networked clients. It offers specialized tools like get_constant for physical constants, get_particle for PDG 2024 properties, and get_material_property for thermophysical or mechanical data. Users can query specific categories including metals, alloys, semiconductors, and ceramics to retrieve properties such as electrical resistivity, thermal conductivity, or refractive index.
Written from the project's README. Every use case is checked against a quote from it.
atomic-context MCP tools 18
Detected from the READMEget_constantsearch_constantslist_constant_categoriesget_particlesearch_particlesget_material_propertysearch_materialslist_material_propertiesget_material_properties_by_categorymetalalloysemiconductor
Show all 18 toolsShow fewer tools
ceramicmechanicalelectricalthermalopticalmagnetic
Configuration
SERVER_PORTSPRING_AI_MCP_SERVER_SSE_ENDPOINT
Automated checks
Ran Nov 14, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachableaskeita/atomic-context is public and reachablegithub.com/askeita/atomic-context ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-13github.com/askeita/atomic-context ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model18 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNo tool named like delete, drop or execPass
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubatomic-context
Read-only MCP server for exact, citable values of fundamental physical constants, particle properties, and material properties — powered by Java Spring Boot + SQLite.
Why this exists
LLMs know physics. Their weights do not reliably reproduce the exact 12-decimal-place CODATA
constants or the specific heat capacity of an obscure alloy at a non-standard temperature.
atomic-context is a grounding layer: a deterministic, citable, offline knowledge server that
any MCP-compatible client (Claude Desktop, Cursor, Continue.dev …) can call.
Data sources (open-source only)
| Dataset | Coverage | License | URL |
|---|---|---|---|
| NIST CODATA 2022 (SRD 121) | 29 fundamental constants | U.S. Gov. Work (public domain) | https://physics.nist.gov/cuu/Constants/ |
| PDG 2024 | 20 particles (leptons, baryons, bosons, quarks, mesons) | CC-BY-4.0 | https://pdg.lbl.gov |
| NIST WebBook SRD 69 | Thermophysical & mechanical properties | U.S. Gov. Work (public domain) | https://webbook.nist.gov |
| AFLOW (aflow.org) | Metals, alloys, intermetallics | CC-BY-4.0 | http://aflow.org |
| OQMD | Metals, alloys, ceramics (DFT) | MIT | http://oqmd.org |
| Materials Project | Semiconductors, ceramics (DFT + exp) | CC-BY-4.0 | https://materialsproject.org |
| NSM Semiconductor Archive | Semiconductors (exp.) | Free academic | https://www.ioffe.ru/SVA/NSM/Semicond/ |
| refractiveindex.info | Optical (refractive index, absorption) | CC0 | https://refractiveindex.info |
MCP Tools
| Tool | Description |
|---|---|
get_constant |
Exact CODATA 2022 value for a physical constant by symbol or name |
search_constants |
Search constants by keyword or category |
list_constant_categories |
List all constant categories |
get_particle |
PDG 2024 properties (mass, spin, charge, lifetime) for a particle |
search_particles |
Search particles by name, symbol, or type |
get_material_property |
Specific property for a material at a given temperature |
search_materials |
Search materials by name, composition, or category |
list_material_properties |
All available properties for a material |
get_material_properties_by_category |
All properties of one category (mechanical/electrical/thermal/optical/magnetic) |
Material categories
- metal — pure metals (Al, Cu, Fe, Ni, Ti, W, Au, Ag, Mg, Co, Mo, Cr)
- alloy — Al alloys (2024/6061/7075), Ti alloys (Ti-6Al-4V, Ti-6242), stainless steels (304, 316L, 17-4PH), carbon steels (1020, 4340), Ni superalloys (Inconel 625/718, Waspaloy, Haynes 230), Cu alloys (brass, bronze, BeCu), Co alloys (Stellite 6), refractory alloys (Mo-TZM), shape memory (NiTi)
- semiconductor — Si, Ge, GaAs, InP, GaN, 4H-SiC, 3C-SiC, GaP, InAs, CdTe, ZnO, Diamond
- ceramic — Al₂O₃, ZrO₂ (8YSZ), Si₃N₄, SiC, B₄C, TiB₂, MgO, BaTiO₃, HA, WC
Property categories
| Category | Properties |
|---|---|
mechanical |
elastic_modulus, tensile_strength_ultimate, yield_strength, hardness_vickers, fracture_toughness, elongation_at_break |
electrical |
electrical_resistivity, carrier_mobility, bandgap, conductivity |
thermal |
thermal_conductivity, specific_heat_capacity, melting_point, thermal_expansion_coefficient, curie_temperature |
optical |
refractive_index, absorption_coefficient, optical_bandgap |
magnetic |
saturation_magnetization, coercivity, relative_permeability, curie_temperature |
Build & run
Requirements
- Java 21
- Maven 3.6+
Build
cd atomic-context
mvn clean package -DskipTests
Transport modes
The server supports two transport modes, selected via a Spring profile.
The @Tool methods are identical in both modes — only the transport layer changes.
stdio — Claude Desktop (local)
Communicates over stdin/stdout. Logging and banner output are disabled, and the
web application type is forced to none to keep the stdio pipe clean.
java -jar target/atomic-context-1.0.0.jar --spring.profiles.active=stdio
Claude Desktop configuration
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS)
or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"atomic-context": {
"command": "java",
"args": [
"-jar",
"/absolute/path/to/atomic-context-1.0.0.jar",
"--spring.profiles.active=stdio"
]
}
}
}
SSE/HTTP — networked clients (default)
Exposes the MCP server over HTTP using Server-Sent Events.
This is the default profile when no --spring.profiles.active flag is provided.
# Default (sse profile)
java -jar target/atomic-context-1.0.0.jar
# Explicit
java -jar target/atomic-context-1.0.0.jar --spring.profiles.active=sse
| Endpoint | URL |
|---|---|
| SSE stream | http://localhost:8080/sse/atomic-context |
| MCP messages | http://localhost:8080/mcp/message |
MCP client configuration
{
"mcpServers": {
"atomic-context": {
"url": "http://localhost:8080/sse/atomic-context"
}
}
}
Production deployment
The port and SSE path can be overridden at runtime without any code changes:
java -jar atomic-context-1.0.0.jar \
--spring.profiles.active=sse \
--server.port=9090 \
--spring.ai.mcp.server.sse-endpoint=/sse/atomic-context
Or via environment variables (Spring Boot maps them automatically):
export SERVER_PORT=9090
export SPRING_AI_MCP_SERVER_SSE_ENDPOINT=/sse/atomic-context
java -jar atomic-context-1.0.0.jar --spring.profiles.active=sse
For HTTPS in production, place a reverse proxy (Nginx, Caddy) in front and proxy to the local port. Critical Nginx settings for SSE:
location /sse/atomic-context {
proxy_pass http://127.0.0.1:8080/sse/atomic-context;
proxy_http_version 1.1;
proxy_set_header Connection "";
proxy_buffering off;
proxy_cache off;
proxy_read_timeout 24h;
}
Configuration files
| File | Purpose |
|---|---|
application.properties |
Shared config: MCP metadata, SQLite datasource, JPA, logging |
application-sse.properties |
SSE transport settings (port, endpoint path) |
application-stdio.properties |
stdio transport + full logging blackout |
Database location
The SQLite database is stored at ~/.atomic-context/atomic-context.db and is seeded
automatically on first startup. The seed scripts in src/main/resources/db/seed/ are
idempotent (INSERT OR IGNORE) and safe to re-run.
License
MIT — data is attributed per-row to original open-source sources.
atomic-context MCP questions
What tools does atomic-context MCP provide?
It exposes 18 tools, including get_constant, search_constants, list_constant_categories, get_particle, search_particles.
What kind of data does atomic-context provide?
It provides exact, citable values for fundamental physical constants, particle properties, and various material properties.
What datasets are used for the information?
The server uses open-source datasets including NIST CODATA 2022, PDG 2024, NIST WebBook, AFLOW, OQMD, Materials Project, NSM Semiconductor Archive, and refractiveindex.info.
Can I use this with Claude Desktop?
Yes, the server supports stdio transport mode which is compatible with Claude Desktop.
Does the server require an internet connection to function?
It acts as an offline knowledge server using a local SQLite database.