Skip to content

docs: add warning and troubleshooting for self-hosted LLM provider config bug (#1577) #1605

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Adityakk9031 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

docs: add warning and troubleshooting for self-hosted LLM provider config bug (#1577) #1605

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 11 additions & 0 deletions apps/docs/self-hosting/configuration.mdx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,6 +39,17 @@ With multiple providers configured, the first one in the order above is used.
Image, video, and high-fidelity PDF understanding require a Gemini or Vertex AI key. Text ingestion, memory extraction, and search work with any provider.
</Note>

> [!WARNING]
> **Known Issue: LLM Provider Configuration Bug (No model provider configured)**
>
> In version `v0.0.8` (and some previous releases), there is a known issue where document processing succeeds at first but the background memory generation fails with `[llm] No model provider configured. Set OPENAI_API_KEY ...`. This occurs even if `supermemory doctor` detects the provider (e.g. `llm provider openai` checks pass) and `OPENAI_API_KEY` is present in the parent process environment.
>
> **Root Cause:**
> The LLM provider registry is instantiated inside a memoized initializer that runs exactly once when the server modules are loaded. Because the background `IngestContentWorkflow` runs in a sandboxed V8 isolate (Miniflare/workerd context), host environment variables on `process.env` are not automatically propagated to the sandbox. Additionally, if the credentials database was decrypted post-startup, the in-memory registry remains unpopulated for the worker.
>
> **Workaround:**
> Ensure that `OPENAI_API_KEY` (or the relevant key for your provider) is exported in the shell environment **before** starting the `supermemory-server` process, and that the server is launched via the generated wrapper script (`~/.local/bin/supermemory-server`) which properly exports keys from `~/.supermemory/env` using `set -a`. If using a process manager (like systemd or PM2) or Docker, explicitly pass the environment variables to the service container/process.

### Fully offline with local models

`OPENAI_API_KEY` + `OPENAI_BASE_URL` covers any OpenAI-compatible endpoint: Ollama, LM Studio, vLLM, llama.cpp server, Together, Fireworks, and more.
Expand Down