Common problems and where to look. When in doubt, read the startup banner — it reports the effective config, every discovered/bridged MCP server, and every failure.

• Is the server still running? The process must stay up for connector discovery and every tool call. If you started it from the wizard's "start now" step, keep that terminal open.
• Native tunnel: does the banner say ready? It reports ready only after /readyz passes and a control-plane poll succeeds. If not, check the runtime key and tunnel ID.
• Right tunnel ID? In ChatGPT the connector must select the same tunnel ID Codexify reports. Set Connection type: Tunnel and Authentication: None (native mode).
• External mode: is your proxy up and pointing at port 3000, with the connector URL ending in /mcp?

• Onboard the chat: ChatGPT Web isn't reliable about reading instructions. Open with "Call get_agent_brief and follow it for the rest of this chat." If it drifts, ask for the brief again.
• Multi-project: project-scoped tools are unavailable until you bind a project. Call set_project_root (exact path) or list_projects (by intent) first. See Multi-Project Mode .
• No repository needed: choose Chat without a project in the setup app, or call set_project_root with withoutProject: true, to get a private scratch workspace. This choice cannot later be changed into a project binding in the same chat.
• import_host_file gone? It's removed when artifactIngress.enabled is false.
• No setup/diff/updater cards? uiWidgets: false deliberately removes Codexify's built-in MCP Apps and their widget metadata. The underlying tools remain available.
• Large bridged tool set not surfacing? Some clients (ChatGPT included) won't show many bridged tools — use gateway mode. See Bridging MCP Servers .

Check the banner first. The most common cause is a wrong command path:

Upstream MCP servers:
 idasql -> FAILED: could not launch 'D:/wrong/path.exe': The system cannot find the path specified.

• A failed upstream is skipped, never blocking startup or native tools.
• disabled: true shows as -> disabled.
• OAuth-only servers need a bearer token via bearerTokenEnvVar or an env-backed Authorization header — OAuth login isn't implemented.
• Legacy SSE/WebSocket transports are rejected explicitly.
• Missing plugin-provided servers? The CLI discovery step may have failed — the banner warns. Pass --codex-cli to make it mandatory, or set codexMcp.useCli.

run_command only runs allowedCommands; exec_command also allows exec.extraAllowedCommands. The allowlist is checked at every command position, and command substitution ($(...), backticks) is rejected outright. Add the binary to the right list, or set exec.mode: "unrestricted" if you fully trust the caller. See Security Model .

That's intentional. Tools stop at a budget and say so on the last line:

(showing lines 1-1000 of 4820 — call again with offset=1000 for the rest)

Call again with the named argument (offset, etc.). Raise the caps in the output block if you must. See Context and Memory .

The shell is picked by name from $SHELL (not the host platform). Starting from Git Bash on Windows gets bash; starting from PowerShell gets PowerShell. Override with exec.defaultShell or a per-call shell argument. Check get_environment to see what's active. See Tools Reference .

• Task state persists in ~/.codexify/projects/<name>-<hash>/memory.json — a new chat picks it up via instructions, or one recall. If memory.enabled is false, nothing is saved.
• Multi-project bindings live in ~/.codexify/conversation-projects/. A conversation can't switch projects — start a new chat. A stale/missing project fails closed. Delete the binding file to clear it.

That's what worktrees are for. In multi-project mode with mode: auto, the second chat gets its own managed worktree. If you set mode: never, they share the checkout. See Worktree Isolation .

Start with the rotating service log:

codexify service logs
codexify service logs -f

• codexify service enable starts an installed service; service install recreates/updates its definition from the selected config.
• self_update accepts only the standard ~/.codexify/bin installation and requires confirm=true after an explicit update request.
• A service-supervised successful update intentionally disconnects MCP while the detached worker swaps the executable and restarts the service.
• After the updated server restarts, open ChatGPT Settings, select the Codexify connector, scroll to the bottom of its tool list, and click Refresh so ChatGPT reloads the updated connector schema.
• With uiWidgets: false, self_update does not emit the updater component or component-only changelog payload; follow progress with codexify service logs -f.
• Windows self-update requires the background service; use the release installer first if Codexify was installed another way.
• A fixed update lock rejects overlapping updates. If an earlier updater really is no longer running, the log identifies the stale lock path and remediation.

• Check for updates performs a fresh release check instead of reusing the normal latest-version cache.
• Doctor runs structured diagnostics asynchronously. Healthy automatic results stay hidden; warnings are summarized, failures expand into individual checks, and Autofix sends the findings to ChatGPT for remediation.
• If setup says the connector schema is stale, use its row-local Refresh action. The app accepts a connector slug only from a same-origin asdk_app_<slug>.web-sandbox.oaiusercontent.com ancestor, then opens the relative #settings/Plugins/plugin_asdk_app_<slug>:~:text=Information-,Refresh,-Connected route through ChatGPT's link-opening API; if that sandbox identity is unavailable it falls back to #settings/Plugins. The app does not send a Refresh prompt to the model or mutate connector state. In Settings, select Codexify if necessary, scroll below the tool list, and click the host-side Refresh control.

Turn up logging:

codexify -v --work-dir /path/to/project # debug
codexify -vv --work-dir /path/to/project # trace
RUST_LOG=codexify=trace,rmcp=warn codexify --work-dir /path/to/project # protocol level

And consider an Audit Logging to see the exact sequence of tool calls (shapes and timings, no payloads).

• FAQ for conceptual questions.
• Open an issue with your startup banner (redact secrets) and what you expected.