v1.0.0-beta.6

Major Release

✨ Knowledge Base is here with File Upload Support!

• With the power of VectorDB's, LLocal now utilizes a persistent way to store vector embeddings. (Although only applicable to File Uploads at the moment).
• pdf, docx, pptx, csv and txt files are supported for file upload.

🌊 What's Changed

• Delete chat is now here. #5
• Fixed the overflow during code generation. #4
• The more button (+) near the text input, has an added file upload section.
• Settings has been revamped now with two sections, Settings (General) and Knowledge Base.
• You can delete the saved VectorDB's through the knowledge base section in settings.
• There's also an auto-complete pop-up for easy navigation through existing VectorDB's by using a / forward slash at the beginning of your prompt.

❄️ Knowledge Base (More information)

• The knowledge base uses a basic similarity search at the moment, which retrieves relevant chunks quickly unlike agentic solutions. Although the results are still not up-to the mark.
• Further, on every RAG based prompt the response also has a sources table attached to show where the retrieved chunks are in the actual file.

Full Changelog: v1.0.0-beta.5...v1.0.0-beta.6

v1.0.0-beta.5

Major Release

✨ Web Search, Image Upload & Linux Support is here!

• Experimental web search is here, you can submit a link as well as perform a duckduckgo search.
• Linux support is here, although the logo of llocal is not visible but the functionalities work just as on the other platforms.
• The auto-install of Ollama works only on debian distros. (Atleast only tested on debian systems)
• Image Upload is now available to be utilized by supported vision models (More Information Here)

🌊 What's Changed

• Copy code block is now available with Syntax Highlight for almost every language out there (with the atom theme haha)!
• Tabular format support is here.
• Stop Generating button is here, although the position is debatable.
• Fixed changing chats while streaming, now you can only switch between chats when the text is not streaming.
• More button (+) in the input for text, which allows to toggle on and off web search aswell as supports images.

❄️ Web Search (More information)

• Web search toggle, has two inbuilt modes a webscraper mode aswell as a duckduckgo search.
• Web scraper : You can simply paste a link in the prompt along with your query and llocal will handle the rest. (Eg. What are the 3 endings in Sekiro? https://en.wikipedia.org/wiki/Sekiro:_Shadows_Die_Twice )
• Web search : This performs a web search, based on a prompt. It returns the top 10 results and a very small summary of each result from the duckduckgo api.

⚡ New Contributors

• @FruitPnchSamuraiG made their first contribution in #3

Full Changelog: v1.0.0-beta.4...v1.0.0-beta.5

1.0.0-beta.4

Major Release

Official Mac Support

• Mac support is here. This has been one of the major goals for the project.
• Ollama, can be downloaded and prompted to be installed if detected as not installed.

What's Changed

• refactor: improve error handling in checkOllama function by @virajbhartiya in #1
• fix : partial mac setup by @dandonarahul2002 in #2

New Contributors

• @virajbhartiya made their first contribution in #1
• @dandonarahul2002 made their first contribution in #2

Full Changelog: v1.0.0-beta.2...v1.0.0-beta.4

1.0.0-beta.3

v1.0.0-beta.3

feat: autoupdater

1.0.0-beta.2

v1.0.0-beta.2

feat: autoupdater

1.0.0-beta.1

Features

Functional Features

• Ollama auto-downloader
• Ollama auto-installer

User experience features

• New logo (I think it might just count as a feature)

1.0.0-beta

Features

Functional Features

• Ollama support (both, pulling models and chatting)
• Local data storage (no cloud connectivity)
• Chat streaming
• Support to switch between models

UX Features:

• Theming options ( across 5 themes)
• Dark mode and light mode support

Note: It is still in a pre-release phase, so all features might not work as intended.