Skip to main content

Cognigy AI agent demo builder MCP server

Project description

cognigy-vibe-mcp

A local Python MCP server for building Cognigy AI agent demos with Claude Code. Provides 15 tools covering the full Cognigy REST API surface — node creation, flow inspection, code push with conflict detection, session testing, and a 20-topic reference library.

Part of the cognigy-claude-plugin.

Installation

uvx cognigy-vibe-mcp

Or install permanently:

uv tool install cognigy-vibe-mcp

Quick start

In your demo project directory, run the cognigy:init-mcp skill in Claude Code. It will:

  1. Create ~/.config/cognigy-mcp/<project-id>/ for state and cache
  2. Create a .cognigy-mcp symlink in your project root
  3. Write the MCP server entry to .claude/mcp.json

Then restart Claude Code and call sync_remote_state to populate state from your Cognigy project.

Manual configuration

Add to .claude/mcp.json:

{
  "mcpServers": {
    "cognigy-vibe": {
      "command": "uvx",
      "args": ["cognigy-vibe-mcp"],
      "env": {
        "COGNIGY_BASE_URL": "https://cognigy-api-au1.nicecxone.com",
        "COGNIGY_API_KEY": "<your-api-key>",
        "COGNIGY_PROJECT_ID": "<your-project-id>"
      }
    }
  }
}

Environment variables

Variable Required Default Description
COGNIGY_BASE_URL Yes Cognigy API base URL
COGNIGY_API_KEY Yes API key from Cognigy UI → My Profile → API Keys
COGNIGY_PROJECT_ID No Default project; prompted if omitted
COGNIGY_VIBE_RESYNC_HOURS No 4 Hours of idle before auto-resync
COGNIGY_VIBE_CACHE_TTL No 300 Resource cache TTL in seconds

Tools

State & sync

Tool Description
sync_remote_state Wipe cache and repopulate from Cognigy remote
get_build_state Return name→ID mappings (filterable by resource_type)
resolve_resource Fast name→ID lookup from local state (no API call)

Flow operations

Tool Description
cognigy_get GET any resource, cache-first
cognigy_list List resources; singular/plural resource_type both accepted
cognigy_create POST resource; extension auto-injected, Say config auto-normalised
cognigy_update PATCH with always-fresh GET + optional deep merge
cognigy_delete DELETE any resource including nodes
cognigy_invoke Named operations: move, clone, train, inject, search
get_flow_chart Chart with relations array + readable hierarchy string

File push

Tool Description
push_code_node Push local .js/.ts to a code node with conflict detection
push_html_node Push local .html to a setHTMLAppState node
push_tool_from_file Create or update a Cognigy tool from a local JSON file

Testing

Tool Description
talk_to_agent Send a message to a flow via REST endpoint

Guidance

Tool Description
explain 20-topic reference library (see below)

explain topics

Call explain() for an overview. Call explain("topic") for full details.

node-positioning · node-wiring · agent-tool-branch · node-config-update · flow-chart-reading · tool-conditions · two-pass-confirm · turn-structure · xapp-delivery · cognigyScript · code-node-patterns · voice-gateway · outbound-trigger · knowledge-store · endpoint-config · function-execution · session-injection · extension-map · node-types · mcp-comparison

Topics are front-loaded in the tool description — no tool call needed to see what's available.

Key behaviours

  • Extension auto-injectioncognigy_create injects extension for all known node types: @cognigy/voicegateway2 for setSessionConfig/hangup, cognigy-ai-agent for AI agent nodes, cxone-utils for xApp nodes
  • Say node normalisationconfig: { text: "Hello" } is automatically wrapped into the full config.say.text envelope
  • Conflict detectionpush_code_node snapshots the last-pushed content and blocks if the Cognigy UI has been edited since
  • Auto-resync — sessions idle longer than the threshold are silently re-synced before the next tool call
  • Atomic writes — state and cache files use write-to-tmp-then-replace to survive process interrupts

Development

git clone https://github.com/ben-elliot-nice/cognigy-claude-plugin
cd cognigy-claude-plugin/cognigy-mcp
uv sync --extra dev
uv run pytest tests/ -v

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

cognigy_vibe_mcp-1.2.6.tar.gz (72.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

cognigy_vibe_mcp-1.2.6-py3-none-any.whl (29.9 kB view details)

Uploaded Python 3

File details

Details for the file cognigy_vibe_mcp-1.2.6.tar.gz.

File metadata

  • Download URL: cognigy_vibe_mcp-1.2.6.tar.gz
  • Upload date:
  • Size: 72.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.19 {"installer":{"name":"uv","version":"0.11.19","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for cognigy_vibe_mcp-1.2.6.tar.gz
Algorithm Hash digest
SHA256 aac8cf32a1ac694dd2cc82154ad279613807d2991cbf57b862c43c29656d0f80
MD5 6840d5432d6dc8520e833e5e4eed4d7c
BLAKE2b-256 36efa5f77f6a1ab9da9b8f01d5294e8067c0e9ef4d1e93b9855b8114569427fc

See more details on using hashes here.

File details

Details for the file cognigy_vibe_mcp-1.2.6-py3-none-any.whl.

File metadata

  • Download URL: cognigy_vibe_mcp-1.2.6-py3-none-any.whl
  • Upload date:
  • Size: 29.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.19 {"installer":{"name":"uv","version":"0.11.19","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for cognigy_vibe_mcp-1.2.6-py3-none-any.whl
Algorithm Hash digest
SHA256 f27a9f6cf33919cebec88e0c0846f71c4078db6c239151a9a95be40005a2eae6
MD5 c88cb8bf0a24de7e7b38ceb891c7a300
BLAKE2b-256 64343fda1e49f72e4c7fc969e216081133e226107e1c2221631e35d1d812e368

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page