Component Gallery
A one to one replica of Loca Client components in various states.
The Client is built with Tauri and Svelte, and both it and this page shares one UI package. A state shown here should be identical to the one in the Client.
If anything looks off, please report it as an issue.
Layouts
Welcome to Loca
Loca runs on a Core: the service that powers your models, speech, and tools.
Welcome to Loca
Welcome to Loca
Loca runs on a Core: the service that powers your models, speech, and tools. Connect this device to one to get started.
Enter the address where your Core is running, then check the connection. You can also press the button on the address bar to find a Core on the same network.
New to this? Read the setup guide.
Connect to a Remote Core
Nearly there. Give this Core a name you'll recognize, then enter the 6-digit pairing code shown on it.
Connect to a Remote Core
Enter the address where your Core is running, then check the connection. You can also press the button on the address bar to find a Core on the same network.
New to this? Read the setup guide.
Connect to a Remote Core
Enter the address where your Core is running, then check the connection. You can also press the button on the address bar to find a Core on the same network.
New to this? Read the setup guide.
Chat
shell wants to run a program| Day | Task | |
|---|---|---|
Sessions
Sessions
Sessions
{
"portion": "1 scoop",
"bowl": "kitchen"
} {
"fed": true,
"dispensedGrams": 45
} only http(s) URLs are allowed
{
"portion": "1 scoop"
} {
"fed": true
} {
"portion": "1 scoop"
} # Today - water the plants - call the vet
HTTP 503 upstream connect error or disconnect/reset before headers
Embedding model timed out after 30s.
Embedding model timed out after 30s.
llama-server: failed to load model: out of memory
whisper model not found on disk
shell wants to run a programmodel wants to use the language modelfetch wants to connect to a serverSettings
Could not reach the Core.
Could not reach the Core.
Whether to automatically deny, or ask you, when a tool requests access this extension never declared.
Whether to automatically deny, or ask you, when a tool requests access this extension never declared.
Whether to automatically deny, or ask you, when a tool requests access this extension never declared.
Whether to automatically deny, or ask you, when a tool requests access this extension never declared.
Could not reach the Core.
~/Documents(required) Open the files you ask it to summarize.api.example.com:443 Look up reference data while drafting.bash command(required) Execute the commands you approve.Enabled, but not offered to the agent while a required permission is denied.
~/Projects(required) Save the edits it makes for you.Runs with full access for the widest compatibility. Turn off to choose permissions yourself.
OAuth sign-in required
Authorize in your browser to connect.
Connection refused: spawn npx ENOENT
Could not reach the Core.
Knowledge
Skill
Knowledge
Knowledge · provided by an extension (read-only)
Knowledge
Knowledge
loca-core serve --port 8787deno run -A --unstable-kv jsr:@loca/core --channel stable --models ~/.loca/modelsLoading...Quick Settings
The essentials to get going. Everything here is also in Settings.
Quick Settings
The essentials to get going. Everything here is also in Settings.
Objects
Searches the codebase for symbols, references, and text matches.
Overlays
Primitives
read_file wants to read a fileThis setting controls how the assistant formats its replies.
Applies on the next message.
The disclosed body content.
Revealed content.
Dialog title
A centered dialog over a dimmed backdrop.
Mobile
llama-server: failed to load model: out of memory
whisper model not found on disk
Bottom sheet
On touch every modal presents as a bottom sheet: a grab handle, drag-to-dismiss, and a slide-up transition, inherited from the one shared primitive.