Skip to main content
Glama

leeshfield MCP

MCP (Model Context Protocol) server for leeshfield (video generation platform). AI clients (Claude, etc.) can log in to a leeshfield account via OAuth, add assets, reference them to generate videos, and receive results.

Architecture

MCP 클라이언트 (Claude 등)
   │  Streamable HTTP + OAuth 2.1 (PKCE)
   ▼
leeshfield-mcp  ← 이 리포. 얇은 어댑터 (리소스 서버)
   │  Bearer 토큰 그대로 전달 (introspection으로 검증)
   ▼
leeshfield 본체  ← 인가 서버(/oauth/authorize, /api/oauth/*) + REST API
  • The authorization server is leeshfield itself. Because the user session (login/consent screen) lives on that domain. This server only advertises the AS location via RFC 9728 protected resource metadata.

  • This server has no authority of its own. Every tool call passes the user access token straight through to the leeshfield API, and workspace permissions/tenant isolation are enforced by leeshfield.

  • Stateless Streamable HTTP — no session state, safe for redeployment and multiple replicas.

Related MCP server: creatify-mcp

Tools

Tool

Description

whoami

Account, active workspace, credit balance

list_models

Model catalog (resolution, aspect ratio, length, attachment limits)

list_assets

Asset list (kind/q/limit filters)

upload_asset

Upload asset via URL or base64 (returns existing asset on duplicate)

estimate_video

Credit estimate and balance check before generation

generate_video

Submit generation job (asset references in attachments, idempotency key automatic)

list_jobs

Status of the latest 50 jobs

get_job

Single job status and result URL (for polling)

Environment Variables

Name

Example

Description

PORT

3000

Listen port

LEESHFIELD_URL

https://leeshfield.klr.kr

leeshfield main server (AS + API)

MCP_PUBLIC_URL

https://leeshfield-mcp.klr.kr

Public URL of this server

OAUTH_INTROSPECT_SECRET

(secret)

Shared secret for leeshfield introspection — must match the main server

Development

npm install
LEESHFIELD_URL=http://localhost:3000 MCP_PUBLIC_URL=http://localhost:3100 \
OAUTH_INTROSPECT_SECRET=dev PORT=3100 npm run dev

Validation: npm run typecheck && npm run lint

Client Connection

# Claude Code
claude mcp add --transport http leeshfield https://leeshfield-mcp.klr.kr/mcp

When connecting, a browser opens and you log in to leeshfield (Google/email), then approve on the consent screen. Client registration (DCR) and token issuance are automatic.

Deployment

  • Branches: develop → dev(dev.leeshfield-mcp.klr.kr, internal network only) / main → prod(leeshfield-mcp.klr.kr)

  • Jenkins → OCI Registry → ArgoCD (gitops apps/leeshfield-mcp)

  • dev MCP points to dev leeshfield(dev.leeshfield.klr.kr), prod points to the prod main server.

F
license - not found
Not graded
quality - not tested
C
maintenance

Maintenance

Maintainers
Response time
Release cycle
Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

View all related MCP servers

Related MCP Connectors

  • MCP server for Google Veo AI video generation

  • MCP server for Wan AI video generation

  • MCP server for Kling AI video generation

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/leestana01/leeshfield_MCP'

If you have feedback or need assistance with the MCP directory API, please join our Discord server