A hidden agent basically allows each individual conversation to maintain custom settings, via an agent that's not exposed to the traditional functionalities allotted for manually created agents (e.g., browsing, maintenance in agents page).
This will be hooked up to the front-end such that any conversation that's initiated with the default agent can then be given custom settings, which in the background creates a hidden agent. This allows us to repurpose all of our existing agents infrastructure for chat-level customization.
Previously emails with url special characters would not get
successfully identified for login. Account creation was fine due to
email being in POST request body. But login with such emails did not
work due to query params not being escaped before being sent to server
This change escapes both the code and email in login URL sent to
server. So login with emails containing special characters like
`email+khoj@gmail.com' works. It fixes both the URL web app sent by
web app directly and the magic link sent to users to their email
This change also fixes accessibility issue of having a DialogTitle in
DialogContent for screen readers.
Resolves#1090
We've been having issues generating diagrams with Excalidraw that are any degree of complexity. By contrast, LLMs are able to handle Mermaid.js syntax a lot better, as it's much more forgiving and has a simpler declarative style. Refer to https://mermaid.js.org/.
Update so that new diagrams are generated with Mermaid.js, while old diagrams generated with Excalidraw can still be viewed.
Fix for #1082 pushed down adding the `data:image/webp;base64' prefix
of the base64 images to the server image gen API. But the code on the
Obsidian and Desktop client still add these prefixes.
This change stops the Desktop, Obsidian clients from adding the prefix
as it is being handled by the API now. It should resolve showing
images inline in those clients as well
Previously we'd use the general openai client, even if the image
generation model has a different api key and base url set.
This change uses the openai config of the image generation models when
set. Otherwise it fallbacks to use the first openai api provider set
This change adds the ability to use OpenAI, Azure OpenAI or any embedding model exposed behind an OpenAI compatible API (like Ollama, LiteLLM, vLLM etc.).
Khoj previously only supported HuggingFace embedding models running locally on device or via HuggingFaceW inference API endpoint. This allows using commercial embedding models to index your content with Khoj.
Previously if the call to this tool AI failed, the API call would
non-gracefully fail on server. This would leave the client hanging in
a wierd state (e.g with spinner running on web app with no indication
of issue).
Also do not show filters in query debug log lines when no filters in query
- Background
Access to the clipboard API is disabled by certain browsers in non
localhost http scenarios for security reasons.
So the copy API key button doesn't work when khoj is self-hosted
with authentication enabled at a non localhost domain
- Change
This change enables copying API keys by manual text highlight + copy
if copy button is disabled
Resolves#1070
