mcpField
D
DiffBeam Gaussian Beam Optics

DiffBeam Gaussian Beam Optics

Remote
github.com/RiaRosenauerScience & Education0Updated 11w ago
Repository

About DiffBeam Gaussian Beam Optics

DiffBeam is a read-only Gaussian beam optics catalog service for AI agents and engineering tools. Its anonymous MCP provides component searches and revision metadata, while workspace analysis and editing are handled separately through a credentialed bridge.

Key features

  1. 01Search fiber catalogs
  2. 02Search fiber collimators
  3. 03Search lens catalogs
  4. 04Read server and catalog revisions
  5. 05Separate credentialed workspace bridge

Use cases

  1. 01Find optical components for a design
  2. 02Check catalog or server revisions
  3. 03Connect an AI to a planner workspace

MCP support

The anonymous, stateless MCP server uses Streamable HTTP and exposes read-only catalog search and server/catalog revision metadata. Workspace operations are available only through a separate credentialed Direct Agent Bridge, not the public MCP.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://catalog-api.gaussian-beam.com/mcp/
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

DiffBeam by Gaussian Beam

DiffBeam is a deterministic Gaussian beam optics service for AI agents and engineering tools. Its anonymous MCP searches optical component catalogs and reports server/catalog revisions. A separate credentialed Direct Agent Bridge analyzes and applies revisioned edits to a user's Gaussian Beam planner workspace.

DiffBeam does not host or proxy a language model. Bring any compatible AI client.

Remote MCP server

MCP clients connect directly with Streamable HTTP. The public endpoint is anonymous, stateless, and read-only. It has no solver, design-link, or workspace authority.

No-install Direct Agent Bridge

An internet-capable AI that can send HTTPS POST requests does not need native MCP installation. The planner's Connect your AI dialog creates a workspace-specific endpoint, temporary scoped credential, and ready-to-copy prompt. The normal flow is:

  1. get_workspace
  2. make one small planner-native operation such as add_lens or set_wavelength
  3. inspect the returned beam feedback and explain the visible revision
  4. refine with the next useful operation; use dry_run only for explicit preview-only work

Read the Direct Agent Bridge guide. Never put a workspace credential in a URL, public page, model configuration, log, or repository.

Public MCP tools

  • get_diffbeam_state
  • search_fibers
  • search_fiber_collimators
  • search_lenses

Credentialed bridge operations

The bridge is not part of the anonymous MCP. It uses the temporary capability from Connect your AI and groups operations into:

  • workspace reads, changes, and revision history;
  • deterministic solving, beam summaries, waist finding, and point measurement;
  • fiber, collimator, and lens catalog searches;
  • planner-native source, element, parameter, unit, and beam-target mutations;
  • undo and revision restore.

Writes require the current expected_revision and create a live planner revision by default. The public MCP cannot access these operations.

Worked optical-design guides

The JSON file in examples/ is a planner-import example for documentation and bridge workflows. It is not an argument accepted by the anonymous catalog MCP.

Repository scope

This public repository contains discovery metadata and integration documentation. It deliberately contains no production credentials, private workspace data, telemetry, or deployment configuration. The documents and examples are licensed under Apache-2.0.

Security issues should be reported privately as described in SECURITY.md.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.