[FIX] Don't get documents from Obsidian client plugin

[beliashou]

Describe the bug

Synchronization of Obsidian vault with Khoj server gets stuck during Force Sync operation. While the connection to the server is established successfully and the initial API calls are made, the synchronization process never completes. The "Updating" moon animation in the Obsidian plugin continues indefinitely without any progress.

Expected behavior: The Force Sync operation should index all markdown files from the Obsidian vault into the Khoj knowledge base, as it successfully did during the previous setup with the same configuration.

Actual behavior: After initiating Force Sync, the server logs show initial connection and setup steps, but the indexing process appears to halt without any error messages or further progress. No documents appear in the knowledge base despite the sync appearing to run (spinning animation).

To Reproduce

Steps to reproduce the behavior:

1. Install Khoj server using Docker on Linux
2. Configure Ollama with llama3.2-vision model as the AI backend
3. Install and enable Khoj plugin in Obsidian
4. Configure server connection in the plugin settings
5. Verify successful connection (connection successful message appears)
6. Click "Force Sync" button
7. Observe that sync process initiates (logs show initial API calls) but never completes
8. Wait several hours - no progress is made, no documents appear in knowledge base

Screenshots

N/A

Platform

• Server:
  • Cloud-Hosted (https://app.khoj.dev)
  • Self-Hosted Docker
  • Self-Hosted Python package
  • Self-Hosted source code
• Client:
  • Obsidian
  • Emacs
  • Desktop app
  • Web browser
  • WhatsApp
• OS:
  • Windows
  • macOS
  • Linux
  • Android
  • iOS

If self-hosted

• Server Version: 1.36.4

Additional context

• Chat functionality works correctly - both through Obsidian plugin and web client
• Previous installation with similar configuration successfully indexed 800+ documents
• Current installation was done after removing Docker volumes from previous setup
• Server logs show successful initial setup and telemetry, but no indexing progress
• No error messages are present in the logs
• Using Ollama with llama3.2-vision model as LLM backend