MCP transport bridge that converts between stdio, SSE, and Streamable HTTP so local MCP clients can reach remote servers, or remote clients can reach local stdio servers.
Open the source and read safety notes before installing.
Source-backed facts for citing this resource, derived directly from the registry — also available as plain text for AI assistants.
Decision playbook
Signals are present but mixed. Use the checklist below to confirm the source and operational safety for your environment.
0
63
—
No baseline selected
No major trust-signal divergence detected in the current selection.
Confirm ownership and provenance before trusting install instructions.
Source link availableRequired
Open the canonical repository and verify ownership.
Source provenance statusRequired
Marked as source-backed.
Metadata reviewed
No reviewed flag detected in metadata.
Validate risk disclosures before installation or API wiring.
Safety notes presentRequired
Review the listed safety guidance before running commands.
Privacy notes presentRequired
Review data handling notes before connecting accounts or secrets.
Trust level risk gateRequired
Trust level does not block evaluation.
Check package metadata and artifact integrity signals.
Install payload available
Install or copy payload is available for review.
Package verification flag
No package verification flag provided.
Checksum metadata
No checksum provided for downloaded artifact.
Use compare context to validate trade-offs before adoption.
Compare tray has multiple entries
Add at least one more entry to compare trust differences.
Baseline comparison available
No baseline peer selected yet.
Diverging trust signals identified
No major trust-signal divergence found.
Setup at a glance
Copy-ready — paste the snippet to get started.
Adoption plan
Current risk score 24/100. Use staged verification before broader rollout.
Validate source and review signals before any execution.
Confirm source provenanceRequired
Source URL/provenance metadata is present.
Confirm metadata review state
No review metadata found; increase manual validation.
Verify install payload
Install/config payload exists and can be inspected.
Confirm safety, privacy, and package integrity signals.
Review safety notesRequired
Safety notes are present.
Review privacy notesRequired
Privacy notes are present.
Verify package integrity metadata
No package verification/checksum metadata.
Adopt in controlled steps based on the selected plan.
Run in isolated sandbox firstRequired
Use a constrained sandbox and observe behavior across multiple tasks.
Roll out graduallyRequired
Roll out to a small cohort before wider usage.
Set monitoring and fallback
Define rollback path and monitor errors after adoption.
Evidence readiness
Missing required evidence: Metadata review. Risk score 31.
Source repository/provenance is listed.
Required in this preset
Review metadata is missing.
Required in this preset
Safety notes are present.
Required in this preset
Privacy notes are present.
Optional in this preset
Package integrity metadata is missing.
Optional in this preset
Install payload is available.
Required in this preset
Required gaps: Metadata review
Decision timeline
Blocking gaps: Check metadata review status. Risk 28.
triage
Source/provenance metadata is available.
triage
Review metadata is missing.
verify
Safety notes are available.
verify
Privacy notes are available.
verify
Package integrity metadata is missing.
rollout
Install payload is available.
Blockers: Check metadata review status
Prerequisite readiness
5 prerequisites to line up before setup. Have accounts and credentials ready first.
Safety & privacy surface
5 safety and 5 privacy notes across 3 risk areas. Review closely: credentials & tokens, network access.
Disclosure: MIT-licensed Python MCP transport bridge. It does not provide domain-specific tools itself; it proxies other MCP servers across transports.
{
"mcpServers": {
"mcp-proxy": {
"command": "mcp-proxy",
"args": ["https://example.com/sse"],
"env": {
"API_ACCESS_TOKEN": "REPLACE_WITH_ACCESS_TOKEN"
}
}
}
}mcp-proxy bridges MCP transports. It can let a stdio-only client such as Claude Desktop connect to a remote SSE or Streamable HTTP server, or expose a local stdio MCP server through SSE or Streamable HTTP for clients that need a network endpoint.
Use it when transport mismatch is the blocker and the downstream MCP server is already trusted. It is especially useful for remote servers, containerized setups, named local servers, and client/server combinations that do not share the same MCP transport.
These sources were reviewed on 2026-06-06. Prefer the live repository, README, PyPI metadata, license, package metadata, example config, CLI entrypoint, MCP server implementation, and proxy server implementation for current setup and behavior details.
API_ACCESS_TOKEN environment variable to
remote servers.Install the PyPI package with uv or pipx:
uv tool install mcp-proxy
pipx install mcp-proxy
For a stdio client connecting to a remote SSE endpoint:
{
"mcpServers": {
"mcp-proxy": {
"command": "mcp-proxy",
"args": ["https://example.com/sse"],
"env": {
"API_ACCESS_TOKEN": "REPLACE_WITH_ACCESS_TOKEN"
}
}
}
}
For exposing a local stdio server on a local network endpoint, bind carefully:
mcp-proxy --host=127.0.0.1 --port=8080 -- uvx mcp-server-fetch
mcp-proxy is only as safe as the endpoints, commands, credentials, and network bindings you give it. Keep local bridges bound to loopback by default, avoid wildcard CORS unless required, and do not expose unauthenticated tool endpoints on shared networks.
Treat proxy configs as sensitive. They can contain bearer tokens, OAuth secrets, server commands, environment variables, endpoint URLs, and downstream MCP server details that should not be committed to shared repositories.
Show that mcp-proxy Transport Bridge is listed on HeyClaude. Paste this Markdown into your README — it renders the badge and links back to this page.
[](https://heyclau.de/entry/mcp/mcp-proxy-transport-bridge)mcp-proxy Transport Bridge side by side with 3 alternatives on trust, install, platform support, and disclosed safety notes — all from reviewed registry metadata.
1 trust signal differ across this comparison (Submitter).
| Field | MCP transport bridge that converts between stdio, SSE, and Streamable HTTP so local MCP clients can reach remote servers, or remote clients can reach local stdio servers. Open dossier | Hosted streamable-HTTP MCP server that exposes inference.sh platform tools for running apps, managing tasks, proxying external MCP connectors, and calling hundreds of hosted AI models from Claude Code, Cursor, and other MCP clients. Open dossier | Built-in Streamable HTTP MCP server for Nuclear Music Player that lets Claude inspect available music-player domains, discover method signatures, describe data types, and control playback, queue, favorites, playlists, dashboard, and provider workflows. Open dossier | Hosted streamable-HTTP MCP server registered as ai.com.mcp/skills-search that exposes Agent Skills Search API tools for discovering skills from the skills.sh registry using the open Agent Skills format. Open dossier |
|---|---|---|---|---|
| Next steps | ||||
| Trust | ||||
| Review status | Not reviewed | Not reviewed | Not reviewed | Not reviewed |
| Package trust | Package not verified | Package not verified | Package not verified | Package not verified |
| Source provenance | Source-backed | Source-backed | Source-backed | Source-backed |
| SubmitterDiffers | oktofeesh1 | kiannidev | oktofeesh1 | kiannidev |
| Install risk | Review first | Review first | Review first | Review first |
| Notes | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ |
| Brand | — | |||
| Category | mcp | mcp | mcp | mcp |
| Source | Source-backed | Source-backed | Source-backed | Source-backed |
| Author | Sergey Parfenyuk | inference.sh | Nuclear | Agent Skills |
| Added | 2026-06-06 | 2026-06-22 | 2026-06-06 | 2026-06-22 |
| Platforms | ||||
| Harness | ||||
| Source repo | — | — | — | — |
| Safety notes | ✓mcp-proxy can expose local stdio MCP servers as network services; keep the host bound to `127.0.0.1` unless remote access is intentional. Passing `--host=0.0.0.0`, permissive CORS, or named-server routes can make tools reachable by other systems on the network. Proxy configuration can include bearer tokens, OAuth client secrets, headers, server commands, environment variables, and working directories. The proxy can spawn arbitrary configured MCP server commands; only use trusted command strings and config files. Remote SSE or Streamable HTTP servers should be authenticated and trusted before forwarding client requests or tool outputs. | ✓The hosted MCP server can execute inference.sh apps, platform tasks, and proxied connector tools that may create, update, or delete data in connected services such as GitHub, Linear, Slack, Notion, or databases. Bearer API keys grant account-scoped access to platform capabilities; rotate compromised keys immediately and avoid committing tokens to repositories or shared MCP config files. MCP proxy features can discover and call tools on remote MCP servers through inference.sh, which expands the runtime trust boundary beyond the primary client configuration. Tool calls may trigger paid inference, connector actions, storage writes, or outbound network requests according to the selected app or connector. Use least-privilege connector authorization and review each proxied server before enabling it in production agent workflows. | ✓Nuclear MCP Server runs inside the local Nuclear desktop app and exposes a Streamable HTTP server on the localhost interface. The `call` tool can execute Nuclear API methods after discovery through `list_methods`, `method_details`, and `describe_type`. Available domains include Queue, Playback, Metadata, Favorites, Playlists, Dashboard, and Providers, so agents can change what is playing and modify local music-player state. Nuclear's plugin and provider system can retrieve streaming sources, metadata, playlists, and dashboard content from third-party services; use providers only where automated access is allowed. Keep the server bound to localhost, avoid exposing the MCP endpoint on a network interface, and require confirmation before letting an agent change playlists, favorites, queues, or provider settings. | ✓Search results expose skill metadata and install instructions; installing a skill can give an agent executable scripts, shell commands, and third-party repository access. Treat discovered skills like unreviewed dependencies—inspect `SKILL.md`, scripts, and publisher sources before `npx skills add` or equivalent installers run. Self-hosted HAPI deployments inherit API rate limits and outbound network access to `https://skills.sh/api`; scope firewall and logging policies accordingly. Hosted remotes process search queries on vendor infrastructure; avoid sending confidential project names or customer identifiers in search strings when policy requires on-prem discovery. |
| Privacy notes | ✓MCP requests, responses, tool outputs, progress events, headers, OAuth tokens, API access tokens, and session identifiers may pass through the proxy process. Named server config files can contain command arguments, environment variables, and credentials for downstream MCP servers. Exposed network endpoints can reveal tool names, server status, and results to clients that can reach the proxy. Logs and troubleshooting output may include endpoint URLs, command names, headers, connection errors, or server names. Store config files outside shared repositories when they include tokens or private server details. | ✓Prompts, files, tool arguments, task metadata, connector payloads, and model outputs are processed by inference.sh and may transit connected third-party MCP services. OAuth-backed connectors can expose account, workspace, issue, message, or repository content to the agent through proxied tool results. API keys, connector tokens, and task logs should be treated as sensitive credentials and kept out of version control and public issue threads. Hosted execution may retain usage, billing, and operational telemetry according to inference.sh policies; review the platform privacy documentation before processing regulated data. | ✓Tool calls and transcripts can include listening history, search terms, artists, albums, track titles, playlist names, favorites, provider choices, dashboard content, and local player settings. The MCP endpoint is local, but connected MCP clients, model providers, logs, screenshots, and shared chat transcripts can still retain music-library and listening-behavior data. Streaming and metadata providers may receive searches, track identifiers, IP addresses, user-agent metadata, or plugin-specific account context according to their own policies. The MCP server URL and port are local connection details; do not publish screenshots or logs that include private player state or provider credentials. | ✓Search queries, skill names, registry metadata, and tool responses enter MCP client context and may be logged by the hosted remote operator. skills.sh registry responses can include publisher URLs, install counts, and repository identifiers that reveal technology choices to shared clients. Installing skills from search results may download repository archives or scripts that process local files; review each skill's privacy posture before use. Self-hosted HAPI CLI mode keeps search traffic on your infrastructure but still contacts the public skills.sh API unless you mirror the backend. |
| Prerequisites |
|
|
|
|
| Install | | | | |
| Config | | | | |
| Citations | ||||
| Claim | Unclaimed | Unclaimed | Unclaimed | Unclaimed |
Connect Claude to inference.sh over streamable HTTP for app execution, task management, connector proxies, and 250+ hosted model and tool endpoints.
Control Nuclear Music Player playback, queues, favorites, playlists, dashboard, and providers through its built-in local MCP server.
Source-backed guides for putting this to work.
Choose MCP stdio, HTTP, SSE, or WebSocket for your deployment.
Loading live community signals…
A short, calm digest of reviewed Claude resources. Unsubscribe any time.