docs: add Lynkr provider guide#12130
Conversation
| --- | ||
| title: "Using Lynkr with Kilo Code" | ||
| description: "Route Kilo Code requests by complexity across local and cloud models with Lynkr, a self-hosted OpenAI-compatible gateway." | ||
| sidebar_label: Lynkr |
There was a problem hiding this comment.
[WARNING]: Add this page to the provider navigation
The docs sidebar is generated from packages/kilo-docs/lib/nav/ai-providers.ts, where Lynkr is still absent. Adding only the overview link leaves the new page out of the Local & Self-Hosted sidebar used to discover and move between provider guides, contrary to the repository's required new-page workflow. Please add /ai-providers/lynkr to that nav section.
Reply with @kilocode-bot fix it to have Kilo Code address this issue.
| ## Configuration in Kilo Code | ||
|
|
||
| 1. **Open Kilo Code Settings:** Click the gear icon in the Kilo Code panel. | ||
| 2. **Select Provider:** Choose "OpenAI Compatible" from the API Provider dropdown. |
There was a problem hiding this comment.
[WARNING]: Document the current custom-provider setup flow
The current Providers tab exposes a Custom provider entry; OpenAI Compatible appears only inside that dialog as the Provider API, not as an API Provider dropdown choice. The dialog also requires Provider ID, Display name, Model ID, and Model Name, so following the listed steps omits required fields and cannot be submitted. Please mirror the current flow documented in openai-compatible.md: connect Custom provider, fill the required identity fields, choose OpenAI Compatible as the Provider API, and then configure the URL and model fields.
Reply with @kilocode-bot fix it to have Kilo Code address this issue.
Code Review SummaryStatus: No Issues Found | Recommendation: Merge Files Reviewed (1 files)
Previous Review Summaries (3 snapshots, latest commit 17e6390)Current summary above is authoritative. Previous snapshots are kept for context only. Previous review (commit 17e6390)Status: 1 Issue Found | Recommendation: Address before merge Overview
Issue Details (click to expand)WARNING
Fix these issues in Kilo Cloud Files Reviewed (1 files)
Previous review (commit afa1b3d)Status: 1 Issue Found | Recommendation: Address before merge Overview
Issue Details (click to expand)WARNING
Fix these issues in Kilo Cloud Files Reviewed (2 files)
Previous review (commit 32416a7)Status: 2 Issues Found | Recommendation: Address before merge Overview
Issue Details (click to expand)WARNING
Fix these issues in Kilo Cloud Files Reviewed (2 files)
Reviewed by gpt-5.6-sol · Input: 45.9K · Output: 4.8K · Cached: 220.8K Review guidance: REVIEW.md from base branch |
|
Both review items addressed: the page now documents the current custom-provider flow (Settings → Providers → Custom provider, with Provider API set to OpenAI Compatible), and the page is added to the generated navigation in |
|
Addressed: the Models step now states the required model name explicitly ( |
| - **Provider API** — select **OpenAI Compatible**. | ||
| - **Base URL** — `http://localhost:8081/v1` | ||
| - **API key** — any non-empty value (provider authentication is handled inside Lynkr's own configuration). | ||
| - **Models** — add the model ID `lynkr-auto` (required). This is a virtual model name: Lynkr's tier routing selects the actual backend model per request. |
There was a problem hiding this comment.
[WARNING]: Include the required model name, not only the model ID
The custom-provider model form has separate required ID and Name fields (CustomProviderValidation.ts rejects an empty name). This instruction still supplies only lynkr-auto as the ID, so following it literally leaves the form invalid and prevents submission. Please also tell users to set a model display name, such as Lynkr Auto.
Reply with @kilocode-bot fix it to have Kilo Code address this issue.
|
Updated: the Models step now includes both the model name (Lynkr Auto) and the model ID ( |
Adds a provider docs page for Lynkr, a self-hosted Apache-2.0 gateway usable with Kilo Code via the OpenAI Compatible provider, plus an entry in the ai-providers index.
Lynkr routes each request by complexity — simple requests to local models (Ollama/llama.cpp/LM Studio), complex ones to a configured cloud provider — and reduces token usage via tool-schema stripping and JSON tool-result compression. Repo: https://github.com/Fast-Editor/Lynkr
Page follows the structure of the existing provider pages (modeled on requesty.md). Happy to adjust to conventions.