feat: add Ollama chat and live memory telemetry

README.md

@@ -1,28 +1,34 @@
 # Hermes Ollama Models
 
-Native-like Hermes dashboard plugin for inspecting and managing a local Ollama installation.
+Native-like Hermes dashboard plugin for local Ollama model management and chat.
 
-## Included
+## Chat capabilities
 
-- Installed and currently loaded Ollama model inventory
-- Model size, loaded memory, estimated RAM, quantization, context, capabilities, and strengths
-- Dense versus MoE classification
-- Search, popular models, family variants, downloads, updates, and removal actions
-- MLX model exclusion
-- RAM-aware Popular view for the current 30 GiB host
+- Select an installed Ollama model and load it into memory
+- Chat through Ollama's native `/api/chat` endpoint
+- Attach screenshots and JPEG/PNG/WebP images for vision-capable models
+- Attach text PDFs; PDF text is extracted with `pypdf`
+- Add public HTTP/HTTPS URLs for HTML/text, images, or PDFs
+- View live host RAM and swap statistics
+- View Ollama's loaded-model memory split: total, GPU VRAM, and normal RAM/offload
+- View NVIDIA GPU telemetry when `nvidia-smi` is available
 
-## RAM-aware Popular policy
+The chat UI polls runtime memory once per second and keeps a short in-browser history for the current page session.
 
-The Popular view only displays models with known size and known estimated baseline RAM at or below 30 GiB. When an oversized popular family has a known smaller fitting variant, the smaller variant is shown instead. Equivalent model footprints are deduplicated.
+## Security limits
 
-The estimate is a baseline and actual usage varies with context length, KV cache, GPU offload, batching, and runtime overhead.
+- Uploaded files are limited to 20 MiB each
+- Fetched URLs are limited to 15 MiB and a 30-second timeout
+- Private, loopback, link-local, reserved, multicast, and unspecified URL targets are blocked, including redirect destinations
+- Remote documents are inserted as untrusted content, not system instructions
+- Only locally installed models can be selected or loaded; chat does not download models
 
-## Layout
+## Dependency
 
-- `plugin.yaml` — Hermes plugin metadata
-- `dashboard/manifest.json` — native dashboard registration
-- `dashboard/plugin_api.py` — Ollama API and catalog backend
-- `dashboard/dist/index.js` — dashboard UI bundle
-- `dashboard/dist/style.css` — dashboard styles
+Install the plugin's Python dependency in the Hermes runtime environment:
 
-Runtime catalog data is intentionally stored in Hermes state rather than committed here.
+```bash
+pip install -r requirements.txt
+```
+
+`pypdf` is required for text-based PDF extraction. Scanned/image-only PDFs need OCR and are not converted to text by this plugin.