Endiagram Mcp

io.github.dushyant30sutharPublished by GitHub account dushyant30sutharv0.3.16Updated Oct 24, 2026

EN Diagram — structural verification for concurrent systems. Pure math, no AI.

VerifiedStreamable HTTPWeb executableDeveloper ToolsData & Analytics

Overview

AI-generated overview

Deterministic graph-theory analysis of concurrent systems written in EN syntax, answering structural, liveness, reachability and equivalence questions and…

What you can do
  • Describe a system's shape, stages, bridges, cycles, critical path, dominators, min-cuts, subsystems, actors and locations, with optional risk findings and per-node centrality
  • Check invariants such as conservation laws, sustainable cycles and depletable sets, including custom rules
  • Test structural liveness and boundedness for deadlock and overflow risks
  • Trace reachability between entities or actions, with path, distance, boundary crossings and guard coverage
  • Compare two systems or apply a patch to one, and merge or extract subsystems
  • Render SVG or PNG diagrams with themes and layer options
When to use it
Useful when you model a concurrent or workflow system in EN syntax and want mathematical answers about deadlocks, bottlenecks, invariants, reachability or whether two designs are structurally the same. Also handy for producing a diagram once the model is settled.
Requirements
Runs as a local npm package via npx or a global install, or as a remote HTTP endpoint. Node.js is needed for the local option. An optional EN_API_URL variable points at the service endpoint. No account or API key is declared.
Before you install
The local package generates a random install ID stored under the user's home directory and sends it as the X-Endiagram-Install-Id header with every request; it can be disabled with ENDIAGRAM_TELEMETRY=off, a telemetry file, or deleting the ID. System models are sent to the remote service. The render tool produces images only when explicitly requested.

Installation

In SourceWeft

  1. Open Endiagram Mcp in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/io.github.dushyant30suthar%2Fendiagram-mcp/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "endiagram": {
      "type": "http",
      "url": "https://api.endiagram.com/mcp"
    }
  }
}

README

@endiagram/mcp

MCP server for EN Diagram — deterministic structural analysis powered by graph theory. Every result is backed by a named mathematical theorem. No AI inside the computation.

Installation

Run directly:

bash
npx @endiagram/mcp

Or install globally:

bash
npm install -g @endiagram/mcp

Connect

Claude Code

bash
claude mcp add endiagram npx @endiagram/mcp

Claude Desktop

Add to your claude_desktop_config.json:

json
{  "mcpServers": {    "endiagram": {      "command": "npx",      "args": ["@endiagram/mcp"]    }  }}

Cursor

Add to .cursor/mcp.json in your project root:

json
{  "mcpServers": {    "endiagram": {      "command": "npx",      "args": ["@endiagram/mcp"]    }  }}

HTTP (zero install)

Any MCP client that supports HTTP transport:

https://api.endiagram.com/mcp

Smithery

bash
smithery mcp add dushyant30suthar/endiagram

Environment Variables

VariableDefaultDescription
EN_API_URLhttps://api.endiagram.comAPI endpoint for the EN Diagram service

Tools

Six questions about any system, plus a render tool. Every tool takes source (EN code or .en/.txt file path). Tool names are shorthand, not specs — read each description before calling; compose and equivalent are mode-based, not general analyzers.

ToolWhat it answersLevers
structureWhat is this system? Shape, stages, bridges, cycles, critical path, dominator tree, min-cuts, subsystems, actors, locations.detect_findings=true flags risks (unguarded-sink, single-cut-path, multi-cut-path); node=X returns per-node centrality (betweenness, closeness, eigenvector).
invariantWhat's always true? Conservation laws, T-invariants (sustainable cycles), depletable sets, deficiency, reversibility.rules (one per line) checks custom claims. Four supported shapes: no bridge that is also hub · every path from X to Y passes through at least one of [A,B,C] (precedence) · no node with centrality above N · removing any single node disconnects at most N others.
liveCan it deadlock? Can entities overflow? Siphons, traps, unbounded cycles, structural liveness and boundedness.—
reachableCan X reach Y? Path, distance, boundary crossings. from/to accept entity or action names.defense_nodes=a,b,c checks whether guards cover every path.
equivalentAre two systems the same, or what changes if I change this one?Compare mode (source_a+source_b): edit distance + spectral cospectrality. Evolve mode (source+patch): plain EN adds; - name removes; same-name replaces.
composeHow do parts combine (merge) or how does a part stand alone (extract)?Merge: source_a+source_b+links (a.entity=b.entity per line). Extract: source+subsystem (names come from structure.subsystems).
renderSVG or PNG diagram. Only call when the user asks to visualize.Themes: Editorial, Primer, Carbon (each ± isDark) or seed-derived from color=#RRGGBB. structure_layers bitmask (1=subsystems, 2=pipelines, 4=cycles, 8=forks, 16=joins, 32=hubs, 64=deadlock, 128=overflow).

EN Syntax

One statement per line:

actor do: action needs: input1, input2 yields: output1, output2

Shared names between yields and needs create connections automatically:

customer do: place order needs: menu yields: orderkitchen do: prepare food needs: order yields: mealwaiter do: deliver needs: meal yields: served customer

Modeling

Same name = same thing. Put all required inputs in one needs: list (AND). Give two actions the same yield-name to offer alternatives (OR). Sequence = one action consuming another's yield. Re-yield stateful resources to keep them persistent; one-shot data stays consumed. Model at the real atomic granularity — split only when the pieces could be accessed independently.

Learn more at endiagram.com.

Telemetry

@endiagram/mcp generates a random install ID on first run, stored at ~/.endiagram/install-id (mode 0600). It is sent with every request as the X-Endiagram-Install-Id HTTP header so we can correlate requests from the same install for debugging issues that the per-IP signal alone cannot track (mobile networks, VPNs, CGNAT all collapse or churn IPs).

No source code, no file paths, no environment variables, and no PII are sent. The install ID is a random opaque UUIDv4 generated locally.

A first-run notice prints to stderr (never stdout — stdout is the MCP JSON-RPC channel) with the disclosure and the opt-out instructions. The notice fires once per install and never again.

Opting out

Any of these three methods disables the install ID:

  1. Set ENDIAGRAM_TELEMETRY=off as an environment variable (also accepts 0, false, no).
  2. Create a file at ~/.endiagram/telemetry containing the word off.
  3. Delete ~/.endiagram/install-id. (A new one is generated on next run unless option 1 or 2 is also set.)

When any of these is active, the X-Endiagram-Install-Id header is not sent at all — the server falls back to its per-IP HMAC cid for correlation, which works fine for short-term per-session tracing.

Full privacy policy: endiagram.com/privacy

License

MIT

Source: README.md at commit 638f7b0

Tools

7

structure

What is this system? Returns shape (Pipeline, Fork-Join, DAG, Star, Cycle, Tree, Complete, etc.), stages with roles, bridge nodes, cycles, parallelism, critical path, dominator tree, min-cuts, subsystems, interface nodes, actors (who does what, workload entropy), locations (where work happens, boundary crossings). Levers: node=X returns per-node centrality (betweenness, closeness, eigenvector) for a specific node. detect_findings=true flags named structural risks — unguarded-sink (sinks reachable via only pipeline actions, no JOIN/HUB gating), single-cut-path (source-sink pairs with only one vertex-disjoint path), multi-cut-path (paths with redundant defense, min-cut > 1). See server instructions for EN language syntax.

Parametersnodesourcerequireddetect_findings

invariant

What's always true — automatic findings and on-demand checks. Automatic outputs: conservationLaws (weighted entity sums constant across executions), sustainableCycles (T-invariants — action sequences returning to start state), depletableSets (entity groups whose simultaneous depletion is irreversible), behavioral.deficiency (0 means structure fully determines dynamics), behavioral.isReversible, behavioral.hasUniqueEquilibrium. On-demand via `rules`: encode domain-specific claims and verify them against the graph — this is how to check things the topology alone can't see (precedence, coverage, centrality bounds, resilience). See the rules parameter for supported sentence shapes.

Parametersrulessourcerequired

live

Can it deadlock? Can entities overflow? isStructurallyLive means every siphon contains a trap — no structural deadlock possible. uncoveredSiphons are entity groups that can drain permanently, with the actors and locations affected. isStructurallyBounded means no entity can accumulate without limit. unboundedCycles are action sequences that could cause overflow. See the server instructions for EN language syntax.

Parameterssourcerequired

reachable

Can X reach Y? Follows directed data flow first; falls back to undirected. `from` and `to` accept entity names or action names (resolved against the program). Path shows each step with actor and location. locationCrossings counts boundary transitions. defense_nodes checks if guards cover all paths; coverage.fullCoverage=false means unguarded routes exist. See the server instructions for EN language syntax.

Parameterstorequiredfromrequiredsourcerequireddefense_nodes

equivalent

Are two systems the same, or what changes if I change this one? Compare mode (source_a + source_b): structural differences, edit distance, spectral equivalence. isCospectral=true means identical graph structure up to relabeling — topologically the same despite different names, actors, or locations. Evolve mode (source + patch): dry-run a change, shows diff plus new/lost bridge nodes. Patch has three directive types — plain EN statement adds an action; a line starting with `-` (and not containing `do:`) removes the named action; a statement whose action name matches an existing one replaces the original. See the server instructions for EN language syntax.

Parameterspatchsourcesource_asource_b

compose

How do parts combine, or how does a part stand alone? Merge mode (source_a + source_b + links): declare which entities in A are the same as entities in B; the combined graph is wired via string-equality of shared names. Extract mode (source + subsystem): pull a named subsystem out as standalone EN with boundary inputs/outputs, actors, and locations. Valid subsystem names come from structure's `subsystems` field — call structure on the source first to discover them. See the server instructions for EN language syntax.

Parameterslinkssourcesource_asource_bsubsystem

render

SVG or PNG diagram. Only call when user explicitly asks to visualize. The rendered image is delivered to the user, not injected into the model's context. See the server instructions for EN language syntax.

ParameterstypeviewcolorthemeisDarkoutputsourcerequiredqualitydirectionstructure_layers

Version history

1
  1. v0.3.16LatestSep 16, 2026

Score

100% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listedRequired15 pts
  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts