@evercam/mcp

v0.0.7

Published

a month ago

MCP server + CLI for Evercam — cameras, AI, copilot, analytics

evercam-mcp

MCP (Model Context Protocol) server + CLI for the Evercam platform.
Talk to your cameras, AI analytics, copilot, and gate reports — from Claude, Cursor, scripts, or any MCP client.

What is this?

@evercam/mcp exposes the full Evercam platform as AI-callable tools via the Model Context Protocol. It wraps three Evercam backends through a single interface:

| Backend | URL | Features | | ------------------- | --------------------- | -------------------------------------------------------- | | evercam-backend | media.evercam.io/v2 | Cameras, projects, snapshots, BIM, timelapses, media hub | | evercam-labs | labs.evercam.io | AI Copilot, presence detection, luminance, heatmaps | | evercam-ai | data.evercam.io | ANPR, gate reports, smart search, site analytics |

Built on @evercam/sdk — no custom HTTP clients, no vendored types.

Installation

Global install (recommended):

npm install -g @evercam/mcp
# or
pnpm add -g @evercam/mcp

Without installing (npx):

npx @evercam/mcp <command>

Note: npx evercam-mcp <command> will not work unless an unscoped evercam-mcp package is published.

From source:

git clone [email protected]:evercam/evercam-mcp.git
cd evercam-mcp
pnpm install
pnpm -C packages/evercam-mcp build

Authentication

evercam-mcp auth login
# Username or email: [email protected]
# Password: ••••••••
# ✓ Logged in successfully (profile: default)

Token is stored securely in ~/.evercam/tokens.json (chmod 0600) and reused on every subsequent command.

For local Claude/ChatGPT MCP usage, EVERCAM_TOKEN is not required after auth login on the same machine.

Note: the SDK currently has no token refresh endpoint; when a token expires, run evercam-mcp auth relogin.

evercam-mcp auth status   # Check current login state
evercam-mcp auth logout   # Clear stored token
evercam-mcp auth relogin  # Re-authenticate and replace stored token

Alternatively, skip the login flow entirely using an environment variable:

EVERCAM_TOKEN=your-token evercam-mcp serve

Use EVERCAM_TOKEN when running in CI, Docker, remote hosts, or shared machines where ~/.evercam/tokens.json is unavailable.

Quickstart (HTTP + REST tools)

# 1) Authenticate once (or set EVERCAM_TOKEN)
npx @evercam/mcp auth login

# 2) Start the HTTP server
npx @evercam/mcp serve --transport http --port 4004

# 3) List tools
curl http://localhost:4004/tools

# 4) Call a tool directly
curl -X POST http://localhost:4004/tools/list-cameras \
  -H "Content-Type: application/json" \
  -d '{}'

CLI Reference

evercam-mcp <command> [options]

`auth`

evercam-mcp auth login [--profile <name>]    # Log in (prompts username + password)
evercam-mcp auth logout [--profile <name>]   # Log out
evercam-mcp auth relogin [--profile <name>]  # Re-authenticate and replace token
evercam-mcp auth status [--profile <name>]   # Show login state + masked token

`profile`

evercam-mcp profile list                     # List all profiles (active marked ●)
evercam-mcp profile add <name> [options]     # Add a new profile
  --base-url <url>                           #   Override backend URL
  --ai-url <url>                             #   Override AI API URL
  --labs-url <url>                           #   Override Labs URL
  --ingest-url <url>                         #   Override Ingest URL
evercam-mcp profile switch <name>            # Switch active profile

`tools`

evercam-mcp tools list                       # Show all tools grouped by domain

`run`

Call any tool directly from the shell. Arguments are passed as key=value pairs.

evercam-mcp run <tool-name> [key=value ...] [--profile <name>]

Examples:

evercam-mcp run list-cameras
evercam-mcp run get-camera cameraExid=roof-cam
evercam-mcp run list-projects
evercam-mcp run get-gate-report-events \
  projectExid=site-abc \
  fromDate=2024-01-01T00:00:00Z \
  toDate=2024-01-31T23:59:59Z
evercam-mcp run list-anpr-events projectExid=site-abc plate=ABC123
evercam-mcp run smart-search projectExid=site-abc labels='["person","car"]'
evercam-mcp run list-copilot-conversations page=1 limit=10
evercam-mcp run get-presence-intervals \
  cameraExid=roof-cam \
  fromDate=2024-01-15T00:00:00Z \
  toDate=2024-01-15T23:59:59Z

`serve`

Start the MCP server.

evercam-mcp serve                            # stdio transport (default)
evercam-mcp serve --transport http           # Streamable HTTP transport
evercam-mcp serve --transport http --port 4004
evercam-mcp serve --profile staging          # Use a specific profile
evercam-mcp serve --allowed-origins https://app.example.com
evercam-mcp serve --timeout 60000            # Tool timeout in ms
evercam-mcp serve --api-key $MCP_API_KEY     # Require ApiKey auth on HTTP endpoints
evercam-mcp serve --allowed-tools list-cameras,get-camera
evercam-mcp serve --denied-tools delete-progress-photo
evercam-mcp serve --enable-rate-limit        # Enable HTTP REST rate limiting (disabled by default)
evercam-mcp serve --enable-rate-limit --rate-limit 120
evercam-mcp serve --enable-mutations         # Allow create/update/delete/action tools

Use with Claude Desktop

1. Log in:

npx @evercam/mcp auth login

2. Add to Claude Desktop config:

Linux: ~/.config/Claude/claude_desktop_config.json
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "evercam": {
      "command": "npx",
      "args": ["-y", "@evercam/mcp", "serve"]
    }
  }
}

Or with an explicit token (useful for CI or shared machines):

{
  "mcpServers": {
    "evercam": {
      "command": "npx",
      "args": ["-y", "@evercam/mcp", "serve"],
      "env": {
        "EVERCAM_TOKEN": "your-token-here"
      }
    }
  }
}

3. Restart Claude Desktop.

You can now ask Claude things like:

"Which cameras are offline right now?"
"Show gate report events for project abc from last week"
"How many vehicles were detected at the main gate yesterday?"
"Create a timelapse for camera roof-cam for all of January"
"Search for persons in project abc between 9am and 5pm today"
"List my last Copilot conversations"
"What days in March had recordings on camera roof-cam?"

Claude will call the right tools, chain results, and reason over them.

Use with Claude Code

1. Log in:

npx @evercam/mcp auth login

2. Add the MCP server via CLI:

claude mcp add evercam -- npx -y @evercam/mcp serve

Or add manually to ~/.claude/settings.json:

{
  "mcpServers": {
    "evercam": {
      "command": "npx",
      "args": ["-y", "@evercam/mcp", "serve"]
    }
  }
}

With an explicit token:

{
  "mcpServers": {
    "evercam": {
      "command": "npx",
      "args": ["-y", "@evercam/mcp", "serve"],
      "env": {
        "EVERCAM_TOKEN": "your-token-here"
      }
    }
  }
}

3. Verify tools are loaded:

claude mcp list
# evercam: npx -y @evercam/mcp serve

Claude Code can now call all Evercam tools while you code — pull real camera exids, fetch live API shapes for type generation, or check gate report data without leaving the terminal.

Use with Codex

1. Log in:

npx @evercam/mcp auth login

2. Add to ~/.codex/config.toml:

[[mcp_servers]]
name = "evercam"
command = "npx"
args = ["-y", "@evercam/mcp", "serve"]

With an explicit token:

[[mcp_servers]]
name = "evercam"
command = "npx"
args = ["-y", "@evercam/mcp", "serve"]

[mcp_servers.env]
EVERCAM_TOKEN = "your-token-here"

Use with Cursor

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

{
  "mcpServers": {
    "evercam": {
      "command": "npx",
      "args": ["-y", "@evercam/mcp", "serve"]
    }
  }
}

While coding, Cursor can:

Pull real camera/project exids to use in tests instead of fake placeholders
Fetch actual API response shapes to generate correct TypeScript types
Verify fixes against real gate report data without leaving the editor

Use with Windsurf

1. Log in:

npx @evercam/mcp auth login

2. Add to Windsurf MCP config:

Global (~/.windsurf/mcp.json) or project-level (.windsurf/mcp.json):

{
  "mcpServers": {
    "evercam": {
      "command": "npx",
      "args": ["-y", "@evercam/mcp", "serve"]
    }
  }
}

3. Reload Windsurf. The Evercam tools will appear in the Cascade agent tool list.

Use with Zed

1. Log in:

npx @evercam/mcp auth login

2. Add to ~/.config/zed/settings.json:

{
  "context_servers": {
    "evercam": {
      "command": {
        "path": "npx",
        "args": ["-y", "@evercam/mcp", "serve"]
      },
      "settings": {}
    }
  }
}

With an explicit token:

{
  "context_servers": {
    "evercam": {
      "command": {
        "path": "npx",
        "args": ["-y", "@evercam/mcp", "serve"],
        "env": {
          "EVERCAM_TOKEN": "your-token-here"
        }
      },
      "settings": {}
    }
  }
}

3. Evercam tools become available as slash commands and in the Agent panel.

Use with Continue.dev

1. Log in:

npx @evercam/mcp auth login

2. Add to ~/.continue/config.yaml:

mcpServers:
  - name: evercam
    command: npx
    args:
      - -y
      - "@evercam/mcp"
      - serve

With an explicit token:

mcpServers:
  - name: evercam
    command: npx
    args:
      - -y
      - "@evercam/mcp"
      - serve
    env:
      EVERCAM_TOKEN: your-token-here

3. Reload VS Code / JetBrains. The @evercam context provider and tools will be available in the Continue chat panel.

Use with Cline

1. Log in:

npx @evercam/mcp auth login

2. Open VS Code → Cline extension → click MCP Servers → Edit MCP Settings.

Add to the settings JSON:

{
  "evercam": {
    "command": "npx",
    "args": ["-y", "@evercam/mcp", "serve"],
    "disabled": false,
    "autoApprove": []
  }
}

With an explicit token:

{
  "evercam": {
    "command": "npx",
    "args": ["-y", "@evercam/mcp", "serve"],
    "env": {
      "EVERCAM_TOKEN": "your-token-here"
    },
    "disabled": false,
    "autoApprove": []
  }
}

3. The server connects automatically. You'll see a green indicator next to evercam in the MCP Servers list.

Use with Goose

1. Log in:

npx @evercam/mcp auth login

2. Add to ~/.config/goose/config.yaml:

extensions:
  evercam:
    name: evercam
    type: stdio
    cmd: npx
    args:
      - -y
      - "@evercam/mcp"
      - serve
    enabled: true

With an explicit token:

extensions:
  evercam:
    name: evercam
    type: stdio
    cmd: npx
    args:
      - -y
      - "@evercam/mcp"
      - serve
    envs:
      EVERCAM_TOKEN: your-token-here
    enabled: true

Or via CLI:

goose mcp add evercam --command "npx -y @evercam/mcp serve"

Streamable HTTP Server

Run as a persistent HTTP server instead of a stdio subprocess. This is useful for:

Remote deployments (server, Docker)
Connecting multiple clients simultaneously
Integrating with evercam-labs or other services over the network

evercam-mcp serve --transport http --port 4004
# Evercam MCP server listening on http://localhost:4004
#   MCP endpoint:     ALL http://localhost:4004/mcp

HTTP endpoints

| Endpoint | Method | Description | | ----------------------- | ------ | ----------------------------------------------- | | /health | GET | Health check { status: "ok", tools: <count> } | | /tools | GET | List all tools (name, description, inputSchema) | | /tools/:toolName | POST | Call a tool directly with JSON args | | /snapshot/:cameraExid | GET | Authenticated latest snapshot proxy | | /mcp | ALL | Streamable HTTP MCP endpoint |

Authorization header: For REST tool calls, you can pass Authorization: Bearer <token> to override the server token for that request.

How it works

Client                              evercam-mcp HTTP server
  │                                          │
  ├── POST /mcp (initialize) ───────────────►│  Creates session
  │◄── sessionId ───────────────────────────┤  Server returns sessionId
  │                                          │
  ├── GET /mcp (SSE stream) ────────────────►│  Opens persistent SSE stream
  │                                          │
  ├── POST /mcp (tool call) ────────────────►│  Send tool call
  │◄── result (via SSE stream) ─────────────┤  Server streams response back

Connect Claude Desktop to a remote server

{
  "mcpServers": {
    "evercam": {
      "url": "http://your-server:4004/mcp"
    }
  }
}

Connect from Node.js (evercam-labs / custom service)

import { Client } from "@modelcontextprotocol/sdk/client/index.js"
import { StreamableHTTPClientTransport } from "@modelcontextprotocol/sdk/client/streamableHttp.js"

const client = new Client({ name: "my-app", version: "1.0.0" })
const transport = new StreamableHTTPClientTransport(new URL("http://localhost:4004/mcp"))
await client.connect(transport)

// Call any evercam-mcp tool
const result = await client.callTool("list-cameras", {})
const cameras = await client.callTool("smart-search", {
  projectExid: "site-abc",
  labels: ["person", "car"],
  fromDate: "2024-01-15T09:00:00Z",
  toDate: "2024-01-15T17:00:00Z",
})

Call a tool without MCP

curl -X POST http://localhost:4004/tools/get-camera \
  -H "Content-Type: application/json" \
  -d '{"cameraExid":"roof-cam"}'

Docker

FROM node:22-alpine
RUN npm install -g @evercam/mcp
EXPOSE 4004
CMD ["evercam-mcp", "serve", "--transport", "http", "--port", "4004"]

docker run -e EVERCAM_TOKEN=your-token -p 4004:4004 evercam-mcp

Multiple environments simultaneously

EVERCAM_TOKEN=$PROD_TOKEN    evercam-mcp serve --transport http --port 4004
EVERCAM_TOKEN=$STAGING_TOKEN evercam-mcp serve --transport http --port 4002

Note: MCP sessions share the server’s base Evercam identity. REST tool calls can override per-request using the Authorization header.

Use as a Node.js Library

Install as a dependency:

pnpm add @evercam/mcp

With a token:

import { createEvercamMcpServer, startHttp } from "@evercam/mcp"

const { server, evercam } = await createEvercamMcpServer({
  credentials: { token: process.env.EVERCAM_TOKEN },
})
await startHttp(server, evercam, 4004)

With username/password:

import { createEvercamMcpServer, startStdio } from "@evercam/mcp"

const { server } = await createEvercamMcpServer({
  credentials: {
    username: "[email protected]",
    password: "your-password",
  },
})
await startStdio(server)

With custom URLs (staging, local dev):

import { createEvercamMcpServer } from "@evercam/mcp"

const { server, evercam } = await createEvercamMcpServer({
  credentials: { token: process.env.EVERCAM_TOKEN },
  baseUrls: {
    baseUrl: "https://staging.evercam.io/v2",
    aiApiUrl: "https://staging-data.evercam.io",
    evercamLabsUrl: "https://staging-labs.evercam.io",
  },
})

All Tools

Cameras

| Tool | Description | Required | Optional | | --------------- | --------------------------- | ------------ | ------------------------------ | | list-cameras | List all accessible cameras | — | projectExid, page, limit | | get-camera | Get camera details | cameraExid | — | | update-camera | Update camera properties | cameraExid | name, isOnline, timezone |

Snapshots

| Tool | Description | Required | Optional | | ----------------------- | ------------------------------------ | ----------------------------- | -------- | | get-live-snapshot-url | Get live snapshot URL | cameraExid | — | | get-snapshot-at | Get snapshot nearest to timestamp | cameraExid, timestamp | — | | get-latest-snapshot | Get most recent snapshot | cameraExid | — | | get-oldest-snapshot | Get oldest snapshot | cameraExid | — | | list-available-days | List days with recordings in a month | cameraExid, year, month | — |

Projects

| Tool | Description | Required | Optional | | ---------------------- | ---------------------------- | ------------- | -------- | | list-projects | List all accessible projects | — | — | | list-project-cameras | List cameras in a project | projectExid | — | | list-project-users | List users in a project | projectExid | — | | list-project-members | List members with roles | projectExid | — |

Media

| Tool | Description | Required | Optional | | ------------------ | ---------------------------- | ------------------------------------------------- | --------------- | | list-timelapses | List timelapses for a camera | cameraExid, fromDate, toDate | — | | create-timelapse | Create a timelapse | projectExid, cameraExid, fromDate, toDate | title | | list-media-hub | List clips in the media hub | projectExid | page, limit |

BIM

| Tool | Description | Required | Optional | | ------------------ | ----------------------------- | ------------------------- | -------- | | list-bim-layers | List BIM layers for a camera | cameraExid | — | | get-bim-snapshot | Get BIM snapshot at timestamp | cameraExid, timestamp | — | | get-bim-model | Get BIM model info | cameraExid | — |

Copilot

| Tool | Description | Required | Optional | | ---------------------------- | ------------------------------- | --------------------------------------- | -------------------------------------------- | | list-copilot-conversations | List AI Copilot conversations | — | page, limit, projectExid, cameraExid | | get-copilot-conversation | Get a conversation by ID | conversationId | — | | list-copilot-messages | List messages in a conversation | conversationId | page, limit | | send-copilot-message | Send a message | conversationId, message | — | | submit-copilot-feedback | Submit message feedback | conversationId, messageId, rating | comment |

rating accepts "positive" or "negative".

Detections

| Tool | Description | Required | Optional | | -------------------------------------- | ---------------------------- | ---------------------------------- | --------------------------- | | get-presence-intervals | Presence detection intervals | cameraExid, fromDate, toDate | — | | get-luminance-readings | Light level readings | cameraExid, fromDate, toDate | — | | get-site-analytics-processing-status | Processing status | — | projectExid, cameraExid | | get-heatmap-intersections | Heatmap intersection data | cameraExid, fromDate, toDate | — |

ANPR / Gate Report

| Tool | Description | Required | Optional | | ------------------------ | -------------------------------- | ----------------------------------- | ------------------------------------------------------------ | | list-anpr-events | License plate recognition events | projectExid | fromDate, toDate, cameraExid, plate, page, limit | | get-gate-report-events | Gate report events | projectExid, fromDate, toDate | cameraExid, page, limit | | list-verified-days | Verified gate report days | projectExid | fromDate, toDate | | list-rois | ROI configurations | projectExid | cameraExid |

Site Analytics

| Tool | Description | Required | Optional | | --------------------------- | ----------------------------------- | ----------------------------------------------------------- | ------------------------------------------------------------------------- | | smart-search | Semantic/visual search over footage | projectExid | cameraExid, query, labels, fromDate, toDate, page, pageSize | | list-smart-search-queries | Saved smart search queries | projectExid | — | | get-detections-heatmap | Detection heatmap data | projectExid, cameraExid, fromDate, toDate, labels | — | | get-gate-report-counts | Event count statistics | projectExid, fromDate, toDate | cameraExid |

Profiles & Multi-environment

Profiles let you switch between production, staging, and local environments without changing env vars.

# Add a staging profile
evercam-mcp profile add staging \
  --base-url https://staging.evercam.io/v2 \
  --ai-url https://staging-data.evercam.io \
  --labs-url https://staging-labs.evercam.io

# Log in on staging
evercam-mcp auth login --profile staging

# List all profiles
evercam-mcp profile list
#   ● default
#     staging

# Switch active profile
evercam-mcp profile switch staging

# Or pass --profile to any command without switching
evercam-mcp run list-cameras --profile staging
evercam-mcp serve --profile staging

Profile config is stored in ~/.evercam/config.json. Tokens per profile in ~/.evercam/tokens.json.

Environment Variables

Environment variables take highest priority — they override both profile config and SDK defaults.

EVERCAM_BASE_URL, EVERCAM_AI_URL, EVERCAM_LABS_URL, and EVERCAM_INGEST_URL are optional. If omitted, the SDK uses production defaults.

| Variable | Default | Description | | -------------------- | ----------------------------- | -------------------------------------- | | EVERCAM_TOKEN | — | Auth token (skips login flow entirely) | | EVERCAM_BASE_URL | https://media.evercam.io/v2 | Evercam backend URL | | EVERCAM_AI_URL | https://data.evercam.io | Evercam AI API URL | | EVERCAM_LABS_URL | https://labs.evercam.io | Evercam Labs URL | | EVERCAM_INGEST_URL | https://ingest.evercam.io | Evercam Ingest URL | | EVERCAM_PROFILE | default | Active profile name |

Copy .env.example to .env — it's loaded automatically on startup.

Architecture

                    ┌─────────────────────────────────────┐
  MCP Client        │           evercam-mcp                │
  (Claude, Cursor,  │                                      │
   custom app)      │  McpServer                           │
       │            │    tools/cameras.ts                  │
       │ MCP        │    tools/snapshots.ts                │
       │ protocol   │    tools/projects.ts                 │
       ▼            │    tools/media.ts                    │
  ┌─────────┐       │    tools/bim.ts                      │
  │  stdio  │       │    tools/copilot.ts                  │
  │   or    │       │    tools/detections.ts               │
  │  HTTP   │       │    tools/anpr.ts                     │
  │ stream │       │    tools/siteAnalytics.ts            │
  └────┬────┘       │         │                            │
       │            │  @evercam/sdk — Evercam class        │
       │            │    evercam.api  ────────────────────►│──► media.evercam.io/v2
       │            │    evercam.ai   ────────────────────►│──► data.evercam.io
       │            │    evercam.axios ───────────────────►│──► labs.evercam.io
       └────────────┘                                      │
                                                           └──────────────────────

Config resolution order

For each setting, priority is:

EVERCAM_TOKEN env var
  → ~/.evercam/tokens.json[activeProfile]
    → @evercam/sdk DEFAULT_BASE_URLS

Auth flow

evercam-mcp auth login
  1. Prompts username + password
  2. new Evercam({ credentials: { username, password } }).connect()
     → POST media.evercam.io/v2/users/login
  3. Stores returned token in ~/.evercam/tokens.json (chmod 0600)

On serve/run:
  1. Reads EVERCAM_TOKEN env var OR ~/.evercam/tokens.json[profile]
  2. new Evercam({ credentials: { token } }); evercam.setToken(token)

Labs API note

EvercamLabsApi from @evercam/api references window.location.href internally (browser-only). Copilot and detection tools bypass this by calling evercam.axios directly with evercam.axios.env.evercamLabsUrl.

Development

git clone [email protected]:evercam/evercam-mcp.git
cd evercam-mcp
pnpm install
pnpm -C packages/evercam-mcp build        # Compile TypeScript → dist/
pnpm -C packages/evercam-mcp dev          # Watch mode (tsx watch)

Adding a new tool

Find the method in evercam-frontend/packages/api/api/
Add to the relevant packages/evercam-mcp/src/tools/<domain>.ts file:

server.tool(
  "tool-name",
  "Human description shown to the AI",
  {
    requiredArg: z.string().describe("What this is"),
    optionalArg: z.number().optional().describe("Page number"),
  },
  async ({ requiredArg, optionalArg }) => {
    const data = await evercam.api.namespace.method(requiredArg, { optionalArg })
    return { content: [{ type: "text" as const, text: JSON.stringify(data, null, 2) }] }
  }
)

Register in packages/evercam-mcp/src/tools/index.ts
Add to TOOL_DESCRIPTIONS in packages/evercam-mcp/src/cli/tools.ts
Run pnpm -C packages/evercam-mcp build — must compile with zero errors

SDK namespace map

| What you want | Use | | ---------------------------------------------- | ----------------------------------------------------------------- | | Cameras, projects, recordings, BIM, media | evercam.api.* | | ANPR, gate report, ROIs, smart search, heatmap | evercam.ai.* | | Copilot, luminance, presence, labs heatmap | evercam.axios.get(\${evercam.axios.env.evercamLabsUrl}/...`)` |

Publishing

# Monorepo release scripts (repo root)
pnpm mcp:publish:patch
pnpm mcp:publish:minor
pnpm mcp:publish:major
pnpm mcp:publish:patch:commit
pnpm mcp:publish:patch:push
pnpm mcp:publish:dry-run

# Or direct:
node ci-scripts/mcp-release.js patch
node ci-scripts/mcp-release.js patch --mutate --commit
node ci-scripts/mcp-release.js patch --mutate --push
node ci-scripts/mcp-release.js patch --dry-run

Release types:

patch: x.y.z -> x.y.(z+1)
minor: x.y.z -> x.(y+1).0
major: x.y.z -> (x+1).0.0

The package prepublishOnly script still runs pnpm -C packages/evercam-mcp build before packaging. Only the dist/ folder is included in the published package.

After publishing:

npm install -g @evercam/mcp   # global install → evercam-mcp command in PATH
npx @evercam/mcp serve        # no install required
# npx evercam-mcp serve        # won't work unless an unscoped package exists

@evercam/sdk — TypeScript SDK this is built on
evercam-frontend — Source of API types
evercam-labs — AI copilot backend
evercam-ai — ML analytics backend
Model Context Protocol — MCP specification

Published

Vulnerabilities

Links

Maintainers

Keywords

Readme

evercam-mcp

What is this?

Table of Contents

Installation

Authentication

Quickstart (HTTP + REST tools)

CLI Reference

auth

profile

tools

run

serve

Use with Claude Desktop

Use with Claude Code

Use with Codex

Use with Cursor

Use with Windsurf

Use with Zed

Use with Continue.dev

Use with Cline

Use with Goose

Streamable HTTP Server

HTTP endpoints

How it works

Connect Claude Desktop to a remote server

Connect from Node.js (evercam-labs / custom service)

Call a tool without MCP

Docker

Multiple environments simultaneously

Use as a Node.js Library

All Tools

Cameras

Snapshots

Projects

Media

BIM

Copilot

Detections

ANPR / Gate Report

Site Analytics

Profiles & Multi-environment

Environment Variables

Architecture

Config resolution order

Auth flow

Labs API note

Development

Adding a new tool

SDK namespace map

Publishing

Related

License

`auth`

`profile`

`tools`

`run`

`serve`