Oracle Readiness MCP Server
Scrapes Oracle Cloud readiness pages and provides MCP tools to search, filter, compare, and report on features across HCM, ERP, SCM, and Service pillars.
README
Oracle Readiness MCP Server — Unified Podman Solution
A single Podman container that:
- Periodically scrapes Oracle Cloud Applications readiness pages (HCM, ERP, SCM, Service, News)
- Persists rich feature data to a SQLite database
- Exposes an MCP server over Streamable HTTP (or stdio)
What's inside
| File | Role |
|---|---|
oracle_scraper.py |
Dual-strategy scraper: docs.oracle.com catalogue (markdownify) + oracle.com what's-new tables (BeautifulSoup). Merges both into a rich feature schema. |
db.py |
SQLite persistence — features, crawl log, document content cache |
server.py |
FastMCP server — 18 tools covering list/search/filter/compare/report/content |
scheduler.py |
Standalone refresh script for cron / sidecar use |
Dockerfile |
Slim Python 3.12 image (OCI-compatible) |
docker-compose.yml |
Single-service podman-compose file with named volume |
MCP Tools provided
| Tool | Description |
|---|---|
list_products |
List tracked Oracle Cloud pillars |
get_cache_status |
Per-product refresh times, entry counts, content cache stats |
list_releases |
All indexed release codes (26C, 26B…) |
list_product_families |
Product families, optionally filtered to a release |
list_modules |
Modules for a release / product family |
get_release_notes |
All entries for a pillar, optionally filtered by release |
search_release_notes |
Full-text search across titles, descriptions, modules |
get_features_by_module |
All features for a release + module |
get_feature_summary |
Statistical summary (impact, setup, AI, Redwood counts) |
get_opt_in_features |
Features requiring Opt-In action |
get_setup_required_features |
Features requiring Setup configuration |
get_high_impact_features |
Large-scale impact features |
get_auto_enabled_features |
Features auto-enabling in a future update |
get_ai_features |
AI / Agent / Generative features |
get_redwood_features |
Features using Oracle Redwood UI |
compare_releases |
Diff two releases for a module (added/changed/removed) |
generate_report |
Filtered report (Pillar/Module/Update/query) with optional full document text |
get_document_content |
Download + cache full text of a readiness document (HTML→MD or PDF text) |
refresh_readiness_data |
Trigger an immediate refresh instead of waiting for the schedule |
ingest_xlsx_dump |
Load features from an Oracle Readiness Reports Centre XLSX JSON dump |
Quick start
# 1. Clone / copy this directory
cd unified-docker
# 2. Copy and (optionally) edit the env file
cp .env.example .env
# 3. Build and start
podman-compose up -d --build
# 4. Check logs
podman-compose logs -f oracle-readiness-mcp
# 5. Test the health endpoint
curl http://localhost:8080/health
Connecting an MCP client
Claude Desktop / Claude Code (stdio via Podman)
{
"mcpServers": {
"oracle-readiness": {
"command": "podman",
"args": ["exec", "-i", "oracle-readiness-mcp", "python", "server.py"]
}
}
}
HTTP (Streamable MCP)
MCP URL: http://localhost:8080/mcp
If READINESS_TOKEN is set:
Authorization: Bearer <your-token>
ICA (Context Studio) Connector Configuration
| Field | Value |
|---|---|
| Name the connection | Oracle Readiness MCP |
| Choose a data source | Others |
| DataSource Type | framer |
| Connection URL | (leave blank — not used for MCP) |
| Connection Token | <your READINESS_TOKEN> (or leave blank if no auth) |
| MCP URL | http://<podman-host>:8080/mcp |
| Re-Ingestion Required | ☐ (unchecked — server keeps its own refresh schedule) |
Replace
<podman-host>with your Podman host IP or hostname. On a local machine:http://localhost:8080/mcp(Podman runs rootless — no special hostname needed).
Data persistence
Data is stored in a named Podman volume oracle-readiness-data at /data inside the container:
/data/readiness.db— SQLite feature database/data/reports/— generated report files (JSON + Markdown)
Manual refresh
# Refresh all products immediately
podman exec oracle-readiness-mcp python scheduler.py
# Refresh only HCM and ERP
podman exec oracle-readiness-mcp python scheduler.py hcm erp
Environment variables
| Variable | Default | Description |
|---|---|---|
READINESS_DATA_DIR |
/data |
Data directory for SQLite DB and reports |
READINESS_REFRESH_HOURS |
6 |
Hours between background refreshes |
READINESS_AUTOSTART_REFRESH |
1 |
Set to 0 to disable background refresh |
READINESS_HTTP_PORT |
8080 |
HTTP port to bind |
READINESS_TOKEN |
(empty) | Bearer token for HTTP auth |
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。