FAQ & Troubleshooting
Find practical solutions and architectural clarification regarding extension compatibility, air-gapped open-source model optimization, operational agent execution modes, and enterprise privacy guarantees.
⚙️General & Extension Compatibility
Q.Can I import my existing VS Code extensions and settings into Morph IDE?
Yes, 100%! Because Morph IDE is built as a native hard fork of Microsoft's core VS Code engine, standard Open VSX and marketplace extensions (such as TypeScript LSP, ESLint, Python Debugger, Prettier, and GitLens) run identically without modification. Upon first launch, an automated wizard imports your existing
settings.json, color themes, and custom keybindings in seconds.Q.Will standard keybindings or Vim extension mappings continue to work normally?
Yes. All core VS Code shortcuts—including
Cmd + P for Quick File Jump and Cmd + Shift + P for the Command Palette—as well as popular extension layouts like VSCodeVim or Neovim function natively without clashing with Morph's native AI commands (Cmd + L for Sidebar Chat and Cmd + K for Inline Quick Edit).⚡Local LLM Configuration & Performance
Q.Why is my local Ollama or LM Studio connection failing or generating tokens slowly?
First, verify that your localized AI daemon is running and explicitly accessible via
http://localhost:11434 (for Ollama) or http://localhost:1234 (for LM Studio). When selecting model identifiers in Morph's AI Provider Settings panel, ensure you specify the exact compiled repository tag (for example, qwen2.5:7b-instruct-q4_K_M rather than simply qwen). For maximum token streaming speeds, confirm that Apple Metal acceleration (macOS) or NVIDIA CUDA GPU layers (Windows/Linux) are enabled in your runtime server.🤖Agent Operational Modes & Guardrails
Q.How should I choose between Normal, Gather, and Agent operational modes?
Use Normal mode for instant conversational answers, syntax clarification, and algorithm drafting without incurring background directory scan overhead. Switch to Gather mode for deep codebase exploration and bug hunting with complete safety—Gather strictly enforces read-only access and never modifies files. Activate Agent mode when you want automated end-to-end multi-file refactoring, background headless terminal test runner execution, and automatic self-correction of compilation stack traces.
Q.How does Morph prevent memory overflows or hallucinations when running noisy terminal scripts?
During autonomous Agent execution turns, Morph enforces built-in token guardrails to protect your LLM context window. Background terminal command outputs are automatically paginated and capped at 20,000 characters, while repository directory structure insertions are capped at 5,000 characters. This guarantees crisp reasoning loops and eliminates token billing exhaustion.
🛡️Privacy & Beta Setup Guidance
Q.Does Morph IDE ever transmit telemetry or source code to proprietary intermediary servers?
Never. Morph IDE enforces an architectural zero-telemetry standard. When utilizing Bring-Your-Own-Key (BYOK) cloud integrations (Anthropic, OpenAI, Google Gemini, Groq), API payload communications route straight from your desktop to official provider endpoints over verified TLS encryption without third-party proxies or billing markups. When utilizing local open-source models (Ollama, LM Studio), your codebase remains 100% offline and air-gapped.
Q.Why do macOS Gatekeeper or Windows SmartScreen warn that the application is unrecognized during installation?
During our initial public beta period, downloadable binary installers are currently unsigned while our corporate developer code signing certificates complete standard OS reputation assessments. You can safely clear Apple Gatekeeper internet download restrictions by running this simple command in your macOS Terminal:
$xattr -cr /Applications/MorphIDE.app
On Windows machines, simply click the underlined text labeled More info in the SmartScreen alert dialog and select Run anyway to open the setup wizard.