@agentdocs1/mcp-server
v1.0.0
Published
UPLO MCP Server — Organizational knowledge for AI agents. 21 tools for searching, querying, and exporting structured organizational context.
Downloads
27
Maintainers
Readme
@agentdocs1/mcp-server
Organizational knowledge for AI agents. 21 MCP tools for searching, querying, and exporting structured organizational context via UPLO.
Quick Start
Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"uplo": {
"command": "npx",
"args": ["-y", "@agentdocs1/mcp-server"],
"env": {
"AGENTDOCS_URL": "https://your-org.uplo.ai",
"API_KEY": "your-api-key"
}
}
}
}Claude Code
claude mcp add uplo -- npx -y @agentdocs1/mcp-serverDocker (self-hosted)
docker run -d \
-e AGENTDOCS_URL=https://your-org.uplo.ai \
-e API_KEY=your-api-key \
-p 3001:3001 \
uplo/mcp-serverTools
| Tool | Description |
|------|-------------|
| search_knowledge | Semantic search across extracted knowledge |
| search_with_context | GraphRAG: vector search + entity resolution + edge traversal |
| search_org_context | Search across organizational context layers |
| get_entry | Retrieve a specific knowledge entry by ID |
| list_entries | List entries with filtering and pagination |
| get_policy | Look up organizational policies |
| get_procedure | Look up standard operating procedures |
| verify_fact | Cross-reference a claim against the knowledge base |
| get_org_chart | Organizational structure and reporting lines |
| find_knowledge_owner | Find subject matter experts for a topic |
| get_org_systems | List organizational systems and integrations |
| get_process_map | Retrieve business process workflows |
| get_org_goals | Active organizational goals and OKRs |
| export_org_context | Full organizational context snapshot |
| get_directives | Strategic priorities and directives |
| export_workspace | Export digital twin as .claude/ workspace |
| get_identity_context | Agent identity and role context |
| log_conversation | Log agent conversation for audit trail |
| propose_update | Suggest knowledge base updates |
| report_knowledge_gap | Flag missing information |
| flag_outdated | Mark stale entries for review |
Features
- 195 extraction schemas across 40 industry packs (legal, healthcare, finance, engineering, and 36 more)
- 1,400+ document formats via Docling + Apache Tika
- GraphRAG search — vector similarity + entity resolution + relationship traversal
- Classification tiers — public/internal/confidential/restricted access control
- Inline DLP — automatic redaction of SSNs, credit cards, API keys, and credentials
- Per-user auth — HMAC-signed MCP tokens with clearance levels
- Audit logging — every tool call logged with timing, user, and result metadata
Environment Variables
| Variable | Required | Description |
|----------|----------|-------------|
| AGENTDOCS_URL | Yes | Your UPLO instance URL |
| API_KEY | Yes | API key (from Settings → API Keys) |
| MCP_TRANSPORT | No | stdio (default) or http |
| MCP_PORT | No | HTTP port (default: 3001) |
| DEFAULT_PACKS | No | Comma-separated schema packs to activate |
Transports
- STDIO (default): For Claude Desktop, Claude Code, Cursor, Windsurf
- HTTP: For self-hosted deployments, Docker, remote access
About UPLO
UPLO is the organizational digital twin for AI. Ingest documents, extract structured knowledge, build org context, and expose it all via MCP. Learn more →
License
MIT
