From 42e7d8b16061ed4c7dd6cdc07052a6b4fb9d9f5f Mon Sep 17 00:00:00 2001
From: Smiral Rashinkar <smiraldr@gmail.com>
Date: Sat, 17 Oct 2026 16:25:55 +0700
Subject: [PATCH] docs: add IO Intelligence provider page

---
 docs/providers/index.json         |  6 +++
 docs/providers/io-intelligence.md | 66 +++++++++++++++++++++++++++++++
 2 files changed, 72 insertions(+)
 create mode 100644 docs/providers/io-intelligence.md

diff --git a/docs/providers/index.json b/docs/providers/index.json
index 31ceaf75..0bddbc04 100644
--- a/docs/providers/index.json
+++ b/docs/providers/index.json
@@ -42,6 +42,12 @@
       "extension": true,
       "cloud": false
     },
+    {
+      "id": "providers/io-intelligence",
+      "title": "IO Intelligence",
+      "extension": true,
+      "cloud": false
+    },
     {
       "id": "providers/kenari",
       "title": "Kenari",
diff --git a/docs/providers/io-intelligence.md b/docs/providers/io-intelligence.md
new file mode 100644
index 00000000..1680c206
--- /dev/null
+++ b/docs/providers/io-intelligence.md
@@ -0,0 +1,66 @@
+---
+sidebar_label: IO Intelligence
+description: Configure IO Intelligence (io.net) in Zoo Code to run open-weight models such as DeepSeek, Qwen, GLM, Kimi and Llama through one OpenAI-compatible API key.
+keywords:
+  - zoo code
+  - io intelligence
+  - io.net
+  - ai provider
+  - open-weight models
+  - deepseek
+  - qwen
+  - llama
+  - api configuration
+---
+
+# Using IO Intelligence With Zoo Code
+
+Zoo Code supports [IO Intelligence](https://io.net/intelligence), the inference API from io.net. One API key gives access to a catalog of open-weight models (DeepSeek, Qwen, GLM, Kimi, Llama, Gemma, gpt-oss and more) over an OpenAI-compatible endpoint.
+
+**Website:** [https://io.net/intelligence](https://io.net/intelligence)
+
+---
+
+## Getting an API Key
+
+1.  **Sign Up/Sign In:** Go to [ai.io.net](https://ai.io.net/ai/api-keys) and create an account or sign in.
+2.  **Create a Key:** Open **API Keys**, create a new key and copy it. The [API keys guide](https://io.net/docs/guides/intelligence/api-keys-and-secrets) covers key management in detail.
+
+---
+
+## Available Models
+
+Zoo Code fetches the model list from the IO Intelligence `/models` endpoint, so the catalog stays current without an extension update. Context window, output limit and per-token prices come from the catalog. The default model is `meta-llama/Llama-3.3-70B-Instruct`.
+
+Two things to know about the list:
+
+- Models that the catalog marks as not supporting tool calling are left out, since Zoo Code depends on tool calls.
+- Some models are limited to higher account tiers. The list is fetched with your key, so it reflects what your account can use.
+
+The full catalog, with tier information, is on the [io.net models page](https://io.net/docs/guides/intelligence/exploring-ai-models).
+
+---
+
+## Configuration in Zoo Code
+
+1.  **Open Zoo Code Settings:** Click the gear icon (<Codicon name="gear" />) in the Zoo Code panel.
+2.  **Select Provider:** Choose "IO Intelligence" from the "API Provider" dropdown.
+3.  **Enter API Key:** Paste your IO Intelligence API key into the "IO Intelligence API Key" field.
+4.  **Select Model:** Choose your desired model from the "Model" dropdown.
+
+---
+
+## Tips and Notes
+
+- **Reasoning models:** Models that return reasoning content, such as the DeepSeek and Kimi thinking variants, stream it to Zoo Code as reasoning output alongside the answer.
+- **Model ids:** The picker uses the same Hugging Face style ids as the io.net API, for example `deepseek-ai/DeepSeek-V3.2` or `Qwen/Qwen3.6-27B`.
+- **One key, many models:** A single key reaches every model your account tier allows, so switching models does not need a new key.
+
+---
+
+## Relevant resources
+
+- [IO Intelligence product page](https://io.net/intelligence)
+- [io.net documentation](https://io.net/docs)
+- [Models on IO Intelligence](https://io.net/docs/guides/intelligence/exploring-ai-models)
+- [API keys and secrets guide](https://io.net/docs/guides/intelligence/api-keys-and-secrets)