Networking issues on macOS
Issues the maintainers labelled both platform:macos and area:networking.
207 issues · 71 open · 136 resolved (66%) · first seen Jan 29, 2026
Is this getting better or worse?
This class of problem is still growing. 112 new reports in the last 90 days vs 89 in the 90 before — +26%. The open backlog peaked at 71 in 2026-08 and sits at 71 today.
Derived by counting each matching issue from its creation date until it closed. First-party
computation over the anthropics/claude-code tracker — the trend is not published
anywhere upstream.
Which builds report this
166 of these reports name the Claude Code build they were running, spanning 95 releases. Heaviest reporters:
- Claude Code v2.1.150
- Claude Code v2.1.233
- Claude Code v2.1.177
- Claude Code v2.1.204
- Claude Code v2.1.156
- Claude Code v2.1.201
- Claude Code v2.1.199
- Claude Code v2.1.251
Report volume, not a defect count: an issue can mention a build without that build having caused the failure. Versions come from the version string in the bug-report template, matched against real published tags.
Do these get fixed?
Median time to close is 36 days across 136 closures. Of the 136 closures with a recorded reason, 22% were closed as completed and 106 as not-planned or duplicate. The most recent completed fix landed in 2026-08.
Workarounds reported by the community
45 of these issues post a workaround someone says worked. The highest-engagement ones:
- [BUG] Remote Control sessions die after ~20 min idle — server TTL ignores keepalives
- …ctivity), idle sessions would survive indefinitely. Current Workaround The only effective mitigation is triggering periodic real model activity (e.g., sending a trivial message via terminal input every ~15 minutes). This…
Found in the description of #32982 · still open - Remote control session fails to resync after connection drop
- If you need a workaround right now, this stale-session / silent-message-loss failure mode is exactly why I built pikiclaw. It runs Claude Code on your own machine but talks to it through Telegram, Feishu, or Weixin inste…
Found in the thread of #28571 · still open - Remote Control (/remote-control) disconnects frequently
- …— it can never recover because every API call returns 400. Workaround: Running Claude Code in tmux and using after drops.
Found in the thread of #33041 · still open - [BUG] Cowork — "Additional allowed domains" allowlist non-functional
- …the sandbox proxy. Would appreciate any update on a fix or workaround short of enabling "All domains". --- Note: English is not my first language. This comment was written with the assistance of an AI(Claude.ai).
Found in the thread of #38984 · still open - Claude Desktop update breaks LAN SSH/network access (OPERON_SANDBOXED_NETWORK=1)
- …there should be a user-facing setting to allow LAN access. Workaround Run Claude Code from the CLI ( command in terminal) instead of Claude Desktop — the CLI does not have the network sandbox.
Found in the description of #37994 · resolved
Excerpts are quoted from the linked issue so you can judge relevance before opening it. Threads that only ask for a workaround, or report that there is none, are excluded. These are community suggestions, not official guidance from Anthropic — read the full thread before applying one.
Matching issues 1–100
Ranked by community engagement (reactions weighted over comments).
- [BUG] Remote Control sessions die after ~20 min idle — server TTL ignores keepalives
- Remote control session fails to resync after connection drop
- Remote Control (/remote-control) disconnects frequently
- [BUG] Cowork — "Additional allowed domains" allowlist non-functional
- [BUG] Allow network egress - All domains' setting not reflected in session proxy JWT
- [BUG] you stupid agent used my real email as user agent string without asking anything
- [Bug] Anthropic API Error: Connection closed mid-response
- Fix: Claude Code 403 "Request not allowed" on macOS with Proxy/VPN (China & Region-restricted Areas)
- [FEATURE] Remote Control: allow localhost proxies (loopback exception or opt-in for ANTHROPIC_BASE_URL=http://127.0.0.1:*)
- Claude Desktop update breaks LAN SSH/network access (OPERON_SANDBOXED_NETWORK=1)
- [BUG] remote-control WebSocket relay keeps reconnecting / commands hang indefinitely on macOS
- [Bug] Persistent API Response Timeout: Retry Loop Since June 22
- [BUG] Intermittent ECONNRESET on both desktop app and standalone CLI, no VPN/proxy
- [BUG] ECONNRESET issues
- [BUG] Support HTTPS proxy (https:// scheme) for HTTPS_PROXY environment variable
- [BUG] ECONNRESET on corporate network - TLS fingerprinting blocks Bun while curl/Node.js work
- [BUG] Frequent UND_ERR_SOCKET connection failures during active sessions
- Remote Control permanently disconnects after rate limit with multiple concurrent sessions
- [BUG] SSL certificate has expired not working since 2.1.190+ on mac
- [BUG] Persistent mid-stream ECONNRESET on large-context sessions (v1/messages?beta=true)
- Frequent ECONNRESET failures with Opus 4.6 (1M context)
- [BUG] Claude Code makes high-frequency version polling requests (~2-4/min) to GCS
- [BUG] Unable to connect to API (ConnectionRefused) on macOS — persists after full reinstall, safe-mode, and reboot (v2.1.204)
- [BUG] Claude Code hangs with Unable to connect to API (ECONNRESET) on every prompt — both CLI and in-app, persists across networks and clean reinstall
- ⎿ API Error: Unable to connect to API...
- [BUG] OAuth error: The socket connection was closed unexpectedly. For more information, pass `verbose: true` in the second argument to fetch()
- Dispatch/Cowork: ECONNRESET when connecting to API
- [BUG] Frequent "API Error: Unable to connect to API (ECONNRESET)" in Claude Code
- Sustained 502 Bad Gateway from mcp-proxy.anthropic.com (Gmail MCP)
- Remote Control: frequent "unknown network error" and session drops when using iOS app
- SSE heartbeat (≤30s) during model reasoning to prevent 'Connection closed mid-response' / stream-watchdog stalls on long (esp. Agent-tool) turns
- [Bug] Frequent tool execution errors with Claude 4.8 Opus model
- [BUG] ECONNRESET on every API call — macOS 26 Tahoe, Darwin 25.4.0, ARM64
- [Bug] Socket connection closed unexpectedly during API communication
- [BUG] ECONNRESET Errors with pro subscription on Ma
- [Bug] Anthropic API Error: Unable to connect to API (ConnectionRefused)
- [Bug] Claude in Chrome WebSocket bridge fails behind Zscaler/corporate TLS proxy (unable to get local issuer certificate)
- [Bug] Anthropic API Error: Unable to connect to API (ECONNRESET)
- [BUG] ECONNRESET errors in session
- [BUG]
- Recurring ECONNRESET during active sessions, exhausting all 10 retry attempts
- [Bug] Remote Control: stale environments cannot be deleted and ghost sessions cause permanent 404 errors
- [BUG] Remote Control bridge registration sent as plaintext HTTP through proxy instead of CONNECT-tunneling
- [BUG] Cowork: all tasks fail with retry loop then ECONNRESET — fails across networks incl. phone hotspot, survived clean reinstall, worked hours earlier same day (macOS 15.7.3, M3)
- [Bug] Frequent ECONNRESET errors with failed retry logic in autonomous sessions
- [BUG] macOS: long sessions die with ECONNRESET — macOS TCP send-buffer default (autosndbufmax=4MB) bursts the growing context upload; one sysctl fixes it
- [Bug] GitHub API requests blocked by proxy egress policy with HTTP 502 in scheduled routines
- MCP sandbox egress: allow <us2.make.com>
- iOS Remote Control: mobile UI stuck on pulsing thinking-icon after `/compact`; stale tool indicators remain; no recovery via app reload
- [Bug] Socket Connection Closed Unexpectedly During API Request
- [BUG] Non-streaming retry receives event-stream response (HTTP/2 stream mixup, Intel Mac)
- [BUG] Unable to connect to API (ECONNRESET)
- [BUG] Persistent intermittent ECONNRESET ("socket closed unexpectedly") on Claude Code — macOS, multi-day, recurring across sessions
- [BUG] Hosted MCP OAuth initiation ignores NODE_EXTRA_CA_CERTS on corporate MITM networks (v2.1.113+)
- [BUG] Remote control disconnects after c.2 hours despite continous interactions
- [BUG] Claude code stopped working
- Voice mode: WebSocket blocked by Cloudflare bot challenge (403 cf-mitigated)
- [BUG] Opus 4.8 (1M) stream hangs after first chunk since 2.1.154; 2.1.153 is last known good
- [BUG] Desktop: Remote SSH session hangs after `Passing plugin(s) to SDK`, no server.spawn RPC ever issued
- [Bug] Tokens consumed without generating output or response
- Background daemon sessions fail permanently on transient socket errors — no retry/reconnect
- Babysit agents stop with 'socket connection closed unexpectedly' API error
- Bug: `tengu_disable_keepalive_on_econnreset` heuristic never triggers on Bun's native fetch errors
- Claude Code API not connecting
- [Bug] API Connection Error: ECONNRESET across all models and sessions
- Não entendo porque pago o plano maior, pra ficar dando erro...
- [BUG] Cowork/Code tab: exit code 126 on VM spawn — missing com.apple.vm.networking entitlement (M5 Max, macOS)
- [BUG] coworkd_upstream_failed connection_refused - persistent across 2 days on Mac
- [BUG] Asking for random login during session
- [Urgent] Widespread ECONNRESET mid-stream — long streamed responses die, short ones survive
- Frequent ECONNRESET on large-context Opus 5 (1M) sessions — RSTs originate from server edge, with SACK gaps on upload
- Remote Control: app messages sync into the session input but never submit (fleet-wide silent message loss); RC never recovers after compaction
- [BUG] Claude Code CLI/Bash tool cannot access LAN devices on macOS — missing Info.plist/NSLocalNetworkUsageDescription, recurring across 5+ closed/locked issues, never fixed
- [BUG] Claude Code loads fixed subset (10) of system CA certificates regardless of CLAUDE_CODE_CERT_STORE value — primary /v1/messages requests fail with UNABLE_TO_GET_ISSUER_CERT
- [BUG] Claude Code doesn't work with `NODE_OPTIONS=--use-system-ca`
- [Bug] Anthropic API Error: ECONNRESET connection reset during message requests with excessive retry loops
- [BUG] Server is Busy & "API Error: Unable to connect to API (ECONNRESET)" errors for 48+ hours in cowork
- Claude Code tab in Desktop App stuck on "Tinkering..." then ECONNRESET
- [BUG] NO_PROXY from managed-settings.json not injected into MCP subprocess environment
- [BUG] Desktop SSH session hangs 3-10 min on "Starting session…" after network-interface change (utun flap)
- A bug with Cowork
- [Bug] Claude Code experiencing severe performance degradation and intermittent failures
- [BUG] Outbound Request to InfluxDB Blocked by Sandbox (2.1.74 -> v2.1.83 -> 2.1.74)
- [BUG] Cowork VM fails to reach API on corporate/institutional network
- MCP: `disabled: true` does not prevent HTTP connection attempts at startup
- [Bug] API Error: Empty or Malformed Response with HTTP 200 Status
- [BUG] Response latency of up to 15 minutes, reproducible in both Desktop app and CLI, fixed only by fully quitting/restarting
- [BUG] ECONNRESET recurs on Claude Code CLI while curl succeeds simultaneously (v2.1.209, arm64 macOS)
- [Bug] Stream stalls immediately on first API request with "Waiting for API response" retry loop
- [BUG] ECONNRESET on home Wi-Fi — curl succeeds at exact same moment Claude Code fails
- 1M-context session becomes unrecoverable: ECONNRESET on large uncached requests, and /compact fails with the same error
- Intermittent ECONNRESET / Connection error on /v1/messages — auth OK, transport-level reset, 3 macOS machines
- Proxy not effective during startup: policy limits fetch runs before configureGlobalAgents
- [BUG] Intermittent "Subprocess initialization did not complete within 60000ms" on macOS — interactive sessions (VS Code extension / desktop app / interactive CLI) hang, `claude -p` works; correlates with TLS-inspection (Norton)
- [BUG] -ECONNRESET
- [BUG] Remote Control: iOS client stuck on "reconnecting" — messages from phone never reach CLI session
- [BUG] Persistent "Unable to connect to API (ConnectionRefused)" on macOS — Claude Code and Claude Cowork both fail while Claude.ai web works
- [Bug] API Error: Socket connection closed unexpectedly
- [Bug] Anthropic API Error: ECONNRESET - Connection Reset During Session
- Claude Desktop 1.3561.0: SSH remote workspaces blocked on LAN — OPERON_SANDBOXED_NETWORK=1 hardcoded with no opt-out