Otherdsh-plugin
dsh-llm-ollama
dsh-llm-ollama is a DeepSeek Harness plugin that adds an Ollama Cloud provider and Web configuration card. Chat uses Ollama's OpenAI-compatible Chat Completions endpoint, while model discovery, web search, and web fetch use Ollama-native APIs. The Web card is owned by dsh-llm-providers-ui, and this package contributes its keyed card and Host route.
Stars
3
Forks
2
Open issues
3
Last push
Sep 24, 2026
Latest release
v0.6.26
24h Growth
+0
7d Growth
+0
Installation
dsh plugin --profile web add --force \Overview
dsh-llm-ollama is a DeepSeek Harness plugin that adds an Ollama Cloud provider and Web configuration card. Chat uses Ollama's OpenAI-compatible Chat Completions endpoint, while model discovery, web search, and web fetch use Ollama-native APIs. The Web card is owned by dsh-llm-providers-ui, and this package contributes its keyed card and Host route.
- Adds an Ollama Cloud provider with model route ollama-cloud and settings namespace llm-ollama.
- Routes chat through Ollama's OpenAI-compatible Chat Completions endpoint via the shared pi-ai-backed adapter.
- Uses Ollama-native APIs for model discovery, web search, and web fetch.
- Adds its keyed Web card under Settings → LLM Providers; dsh-llm-providers-ui owns the page and shared order store.
- Fetches available models through authenticated Connection RPC, reads /api/tags, deduplicates ids, and enriches up to six models via /api/show.
- Displays Cloud usage windows reported by GET <baseURL>/usage as consumed-percentage meters plus per-model request counts.
- Supports a sortable model catalog with context window, Default thinking, and capability flags in row details.
- Registers ollama-cloud web_search and web_fetch providers; deployment policy is set in the profile patch.