feat: add adaptive Ollama catalog filtering

README.md

@@ -18,7 +18,25 @@ Native-like Hermes dashboard plugin for local Ollama model management and chat.
 - Streamed Ollama responses with a real Stop action that cancels the active request
 - Minimized-by-default expandable thinking/progress details with live stage, elapsed time, event, and character counters
 
-The chat transcript and selected model are persisted in this browser, so navigating away from the plugin or reloading the dashboard does not clear the conversation. Use **Clear chat** to remove the saved transcript. The UI also shows a prominent live processing status while Ollama is working, including an animated indicator, elapsed time, request preparation, and response-generation stages. This is operational progress only; private model chain-of-thought is not exposed. Uploaded files remain temporary and are not stored in browser persistence.
+The chat transcript and selected model are persisted in the Hermes server's SQLite database at `~/.hermes/ollama-manager/chat.sqlite3`, so conversations can be listed and resumed from another browser or after a dashboard restart. Use **New conversation** to start a separate thread and **Clear chat** to delete the selected shared conversation. Uploaded files remain temporary; attachment names/types/URLs are retained as metadata, not raw file contents.
+
+## Performance metrics
+
+Each completed request records Ollama-provided counters and timing when available:
+
+- Time to first token (TTFT)
+- Total request latency
+- Prompt/input token count and prompt tokens/sec
+- Output token count and output tokens/sec
+- Total tokens
+- Ollama total/load/prompt-evaluation/generation durations
+- Model, request status, request ID, and error details
+
+The dashboard shows per-conversation metrics and rolling aggregates across the latest 100 requests. Values are shown as unavailable when Ollama does not provide them; the plugin does not estimate token counts.
+
+## Catalog and download filters
+
+The Popular and Available downloads views only show models with known size and expected RAM estimates that fit the running host's installed RAM, detected from `/proc/meminfo` and reported in the API filter metadata. Available downloads also includes known fitting variants from installed model families. The dashboard can filter downloads by Dense/MoE type or ability (completion, thinking, tools, vision, audio, or video), and organize them by popularity, newest, size, or name.
 
 ## Security limits