docs: add Cohere Parse v5 OCR support for cohere and azure_ai

docs/ocr.md

@@ -5,12 +5,14 @@
 | Cost Tracking | ✅ |
 | Logging | ✅ (Basic Logging not supported) |
 | Load Balancing | ✅ |
-| Supported Providers | `mistral`, `azure_ai`, `vertex_ai` |
+| Supported Providers | `mistral`, `azure_ai`, `vertex_ai`, `cohere` |
 
 :::tip
 
 LiteLLM follows the [Mistral API request/response for the OCR API](https://docs.mistral.ai/capabilities/vision/#optical-character-recognition-ocr)
 
+The Cohere Parse OCR integration supports image inputs only
+
 :::
 
 ## **LiteLLM Python SDK Usage**
@@ -347,4 +349,4 @@ The response follows Mistral's OCR format with the following structure:
 | Mistral AI  |   [Usage](#quick-start)                 |
 | Azure AI    |   [Usage](../docs/providers/azure_ocr)                 |
 | Vertex AI   |   [Usage](../docs/providers/vertex_ocr)                 |
-
+| Cohere      |   [Usage](../docs/providers/cohere#parse-ocr)                 |

docs/providers/azure_ocr.md

@@ -1,4 +1,4 @@
-# Azure AI OCR (Mistral)
+# Azure AI OCR
 
 ## Overview
 
@@ -146,9 +146,51 @@ response = await litellm.aocr(
 Azure AI OCR endpoints don't have internet access. LiteLLM automatically converts public URLs to base64 data URIs before sending requests to Azure AI.
 :::
 
+## Cohere Parse
+
+LiteLLM supports Cohere Parse v5 through Azure AI Foundry with the `azure_ai/Cohere-parse-v5` route. Set `AZURE_AI_API_KEY` and `AZURE_AI_API_BASE`, for example `https://<resource>.services.ai.azure.com/models`. LiteLLM rewrites this base URL to `/providers/cohere/v2/parse`
+
+### LiteLLM SDK
+
+```python showLineNumbers title="Cohere Parse SDK Usage"
+import litellm
+import os
+
+os.environ["AZURE_AI_API_KEY"] = "your Azure AI key"
+os.environ["AZURE_AI_API_BASE"] = "https://<resource>.services.ai.azure.com/models"
+
+response = litellm.ocr(
+    model="azure_ai/Cohere-parse-v5",
+    document={
+        "type": "image_url",
+        "image_url": "https://example.com/image.png",
+    },
+    output_format="markdown",
+)
+```
+
+### LiteLLM Proxy
+
+```yaml showLineNumbers title="proxy_config.yaml"
+model_list:
+  - model_name: cohere-parse-v5
+    litellm_params:
+      model: azure_ai/Cohere-parse-v5
+      api_key: "os.environ/AZURE_AI_API_KEY"
+      api_base: "os.environ/AZURE_AI_API_BASE"
+    model_info:
+      mode: ocr
+```
+
+The `azure_ai` provider fronts several OCR backends and picks one from the model name in `litellm_params.model`. Any name containing both `cohere` and `parse`, case-insensitively, routes to Cohere Parse, so a custom Foundry deployment name works as long as it contains both strings. A name missing either string falls through to the Mistral OCR route and fails against a Cohere endpoint, so keep `litellm_params.model` as `azure_ai/Cohere-parse-v5` if your deployment is named differently; the `model_name` alias can be anything
+
+Cohere Parse accepts image inputs only. Use an `image_url` document with an image URL or a base64 image data URI. `document_url` and PDF inputs raise an error. Local image files passed with `{"type": "file", ...}` are converted to data URIs by LiteLLM
+
+Foundry cannot fetch external image URLs, so LiteLLM downloads remote images and inlines them as base64 data URIs before sending the request. The response uses the standard OCR format with `pages[].index`, `pages[].markdown`, `pages[].images`, and `usage_info.pages_processed`
+
 ## Supported Models
 
 - `mistral-document-ai-2505` - Latest Mistral OCR model on Azure AI
+- `Cohere-parse-v5` - Cohere Parse v5 OCR model on Azure AI
 
 Use the Azure AI provider prefix: `azure_ai/<model-name>`
-

docs/providers/cohere.md

@@ -207,6 +207,74 @@ print(response)
 </TabItem>
 </Tabs>
 
+## Parse (OCR)
+
+LiteLLM supports Cohere Parse v5 for image OCR through the direct Cohere route `cohere/parse-v5.0`. The direct route uses `COHERE_API_KEY`, sends requests to `https://api.cohere.com/v2/parse`, and is available through `litellm.ocr()`, `litellm.aocr()`, and the proxy OCR endpoint
+
+<Tabs>
+<TabItem value="sdk" label="LiteLLM SDK Usage">
+
+```python
+import litellm
+import os
+
+os.environ["COHERE_API_KEY"] = "cohere key"
+
+response = litellm.ocr(
+    model="cohere/parse-v5.0",
+    document={
+        "type": "image_url",
+        "image_url": "https://example.com/image.png",
+    },
+    output_format="markdown",
+)
+
+for page in response.pages:
+    print(page.index)
+    print(page.markdown)
+```
+
+Set `output_format="blocks"` to include `pages[].blocks` in the standard `OCRResponse`. The default output format is `markdown`. Set `req_format="native"` to return the raw Cohere response
+
+</TabItem>
+
+<TabItem value="proxy" label="LiteLLM Proxy Usage">
+
+Add this to your LiteLLM proxy config.yaml
+
+```yaml
+model_list:
+  - model_name: cohere-parse-v5
+    litellm_params:
+      model: cohere/parse-v5.0
+      api_key: os.environ/COHERE_API_KEY
+    model_info:
+      mode: ocr
+```
+
+Call the proxy at `/v1/ocr` or `/ocr`
+
+```bash
+curl http://0.0.0.0:4000/v1/ocr \
+  -H "Authorization: Bearer sk-1234" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "cohere-parse-v5",
+    "document": {
+      "type": "image_url",
+      "image_url": "https://example.com/image.png"
+    },
+    "output_format": "markdown"
+  }'
+```
+
+</TabItem>
+</Tabs>
+
+Cohere Parse accepts images only. Use an `image_url` document with an image URL or a base64 image data URI. `document_url` and PDF inputs raise an error. Local image files passed with `{"type": "file", ...}` are converted to data URIs by LiteLLM
+
+The standard OCR response includes `pages[].index`, `pages[].markdown`, `pages[].images`, and `usage_info.pages_processed`. Cohere Parse costs $0.0015 per page
+
 
 ## Supported Models
 | Model Name | Function Call |