/CHANGE_LOG.md

23-10-2025:

• Fixed Ollama cloud model support with new OllamaChat node
• Cloud models authenticate via SSH keys (no API key required)
• Uses ollama.chat() method for cloud compatibility

04-08-2025:

• Added think option and separate thinking output on the Ollama Generate node. This replaces the previous filter_thinking workaround.
• Added tooltips and node descriptions.
• Added example workflows. See ComfyUI's template browser in the workflow menu.
• Added a message that appears when the list of models fails to load. User can click the "Reconnect" button on the connectivity node to reload list.
• Deprecate V1 nodes, please replace them in your workflows.

10-07-2025:

Adding user seed to OllamaVision request.

10-06-2025:

Increased temperature limits in OptionsV2.

26-05-2025:

Added toggle to strip thinking from response

22-04-2025:

Fix v2 settings num ctx limit. Update ollama python 0.4.8.

10-01-2025:

V2 release.

Added nodes:

• OllamaGenerateV2

• OllamaConnectivityV2

• OllamaOptionsV2

09-01-2025

Update python ollama client 0.4.2 -> 0.4.5

04-12-2024:

Fix image base64 to binary issue.

25-11-2024:

Fix some bug in fetching model names.

20-11-2024:

Support step of 1 in keep_alive.

16-11-2024:

Support -1 value in keep_alive.

19-10-2024:

Add seed input to OllamaVision node.

09-10-2024:

Add out put format text or json in all generate nodes.

(should be compatible with existing workflows)

31-08-2024:

Merge change top_k and num_predict params in OllamaAdvanced to be INT instead FLOAT.

30-07-2024

Add 2 new Nodes:

• OllamaSaveContext: will save context from OllamaGenerateAdvance to a png file in saved_context folder.
• OllamaLoadContext: will load context for OllamaGenerateAdvance from a png files in saved_context folder.
• context input in OllamaGenerateAdvance node now can handle string also.

18-06-2024

• Add keep_context option in OllamaGenerateAdvance, this flag will preserve context from the current run in the node itself.
• Modify debug flag to be boolean.

24-05-2024

• Add drop down list to select ollama models that currently available in ollama server.

23-05-2024

• Add keep_alive option. Controls for how long the selected model will stay loaded into memory following the request (default: 5m).

21-05-2024

• Modify Ollama Generate Advance node, change seed to 'INT' from 'FLOAT'.