Prompt Forge MCP Server
Python MCP server that exposes Prompt Forge secure gateway tools over the Model Context Protocol.
Works with any MCP-compatible AI CLI: Claude Code, Claude Desktop, Gemini CLI, OpenAI Codex CLI, or any custom MCP client.
Install
pip install promptforge-mcp
Quick Start
Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"promptforge": {
"command": "promptforge-mcp",
"env": {
"PROMPTFORGE_API_KEY": "your-api-key",
"PROMPTFORGE_ORG_ID": "your-org-id",
"PROMPTFORGE_ADMIN_TOKEN": "your-admin-token"
}
}
}
}
Claude Code
claude mcp add promptforge -- promptforge-mcp
Then export the required environment variables in your shell before launching Claude Code.
Gemini CLI
Add to ~/.gemini/settings.json:
{
"mcpServers": {
"promptforge": {
"command": "promptforge-mcp",
"args": [],
"env": {
"PROMPTFORGE_API_KEY": "your-api-key",
"PROMPTFORGE_ORG_ID": "your-org-id",
"PROMPTFORGE_ADMIN_TOKEN": "your-admin-token"
}
}
}
}
OpenAI Codex CLI
Add to ~/.codex/config.json:
{
"mcpServers": {
"promptforge": {
"command": "promptforge-mcp",
"args": [],
"env": {
"PROMPTFORGE_API_KEY": "your-api-key",
"PROMPTFORGE_ORG_ID": "your-org-id",
"PROMPTFORGE_ADMIN_TOKEN": "your-admin-token"
}
}
}
}
Run directly (stdio)
Any MCP client that launches a subprocess can use this directly:
export PROMPTFORGE_API_KEY="your-api-key"
export PROMPTFORGE_ORG_ID="your-org-id"
export PROMPTFORGE_ADMIN_TOKEN="your-admin-token"
promptforge-mcp
Run as HTTP/SSE server
For clients that connect over HTTP instead of spawning a subprocess:
promptforge-mcp --http
# or
MCP_TRANSPORT=http MCP_PORT=8765 promptforge-mcp
Then point your MCP client at http://localhost:8765/sse.
Environment Variables
| Variable | Required | Default | Description |
|---|---|---|---|
PROMPTFORGE_API_KEY |
Yes (for gateway) | — | API key for gateway requests |
PROMPTFORGE_ADMIN_TOKEN |
Yes (for admin tools) | — | JWT token for secure mode, billing, providers, and audit tools |
PROMPTFORGE_ORG_ID |
Yes (for admin tools) | — | Default organization ID |
PROMPTFORGE_API_BASE |
No | https://api.secureprompt.tech |
API base URL |
PROMPTFORGE_GATEWAY_URL |
No | {API_BASE}/v1/gateway/requests |
Gateway endpoint override |
PROMPTFORGE_TIMEOUT |
No | 30 |
Request timeout in seconds |
PROMPTFORGE_MCP_TOOLS_PATH |
No | — | Override path for tool definitions JSON |
MCP_RATE_LIMIT_PER_MINUTE |
No | 60 |
Rate limit per minute |
MCP_TRANSPORT |
No | stdio |
Transport mode (stdio or http) |
MCP_HOST |
No | 0.0.0.0 |
SSE server host |
MCP_PORT |
No | 8765 |
SSE server port |
MCP Tools
Gateway
| Tool | Description |
|---|---|
promptforge.secure_generate |
Send a request through the secure LLM gateway |
promptforge.list_gateway_requests |
List gateway request logs for an org (metadata only) |
Secure Mode
| Tool | Description |
|---|---|
promptforge.redact_and_tokenize |
Tokenize sensitive entities in text |
promptforge.detokenize |
Restore tokenized text using a vault reference |
promptforge.list_secure_mode_entities |
List valid entity types for redaction |
promptforge.get_secure_mode_config |
Get org's Secure Mode configuration |
Policies
| Tool | Description |
|---|---|
promptforge.policy_dry_run |
Evaluate a policy decision without executing a call |
Audit
| Tool | Description |
|---|---|
promptforge.audit_query |
Query audit logs (admin-only) |
Billing
| Tool | Description |
|---|---|
promptforge.get_billing_status |
Get billing status and usage metrics for an org |
promptforge.list_billing_plans |
List available billing plans |
Providers
| Tool | Description |
|---|---|
promptforge.list_providers |
List configured LLM providers for an org |
promptforge.test_provider |
Test a provider connection and validate credentials |
Docker
docker build -t promptforge-mcp .
docker run -p 8765:8765 \
-e PROMPTFORGE_API_KEY=your-key \
-e PROMPTFORGE_ORG_ID=your-org \
-e PROMPTFORGE_ADMIN_TOKEN=your-token \
promptforge-mcp --http
License
MIT
Metadata
Release files for promptforge-mcp 0.2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| promptforge_mcp-0.2.0.tar.gz | 8.9 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| promptforge_mcp-0.2.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 19.0 kB
Release files / promptforge_mcp-0.2.0.tar.gz
| Download URL | promptforge_mcp-0.2.0.tar.gz |
|---|---|
| Size | 8.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
62cb1b19e08f4c52801d8dcd9d8d1c1d4449b0e22d4702950fd27c3312131e8e
|
|
BLAKE2b-256 checksum How to use checksums |
fb1efb5df404738b97cf135c2860ee6eb8672dbc00a5a1a3733a379130533119
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.4
|
Release files / promptforge_mcp-0.2.0-py3-none-any.whl
| Download URL | promptforge_mcp-0.2.0-py3-none-any.whl |
|---|---|
| Size | 10.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
6b52fbe4a2ff22f343f76f5311ae02ac987a0e7f1552198206502d59780a2743
|
|
BLAKE2b-256 checksum How to use checksums |
5ab40e0d27f71d7e9dea8dd64faab15f53e7ea5af8643e2b00d999146ace6110
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.4
|
Security for promptforge-mcp 0.2.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.