summary system backend implementation
This commit is contained in:
@@ -116,11 +116,9 @@ all projects use isolated memory. Returns `201` with the created project object.
|
||||
| `icon` | string | Icon identifier |
|
||||
| `isolated` | integer | Memory isolation flag (always 1) |
|
||||
| `notes` | string | User-authored project notes |
|
||||
| `system_prompt` | string | Per-project system prompt override (null = use global) |
|
||||
|
||||
Only provided fields are updated — omitted fields are not touched. This
|
||||
enables safe partial updates (e.g. saving just `notes` without affecting
|
||||
`name` or `colour`). Both orchestration and memory service implement dynamic
|
||||
field patching.
|
||||
Only provided fields are updated — omitted fields are not touched.
|
||||
|
||||
### Models
|
||||
|
||||
@@ -161,7 +159,8 @@ Returns `503` if llama-server is unreachable.
|
||||
"temperature": 0.65,
|
||||
"repeatPenalty": 1.3,
|
||||
"topP": 0.9,
|
||||
"topK": 41
|
||||
"topK": 41,
|
||||
"systemPrompt": "You are a helpful assistant..."
|
||||
}
|
||||
```
|
||||
|
||||
@@ -177,6 +176,7 @@ Returns `503` if llama-server is unreachable.
|
||||
| `repeatPenalty` | float | 1–2 | Repeat token penalty |
|
||||
| `topP` | float | 0–1 | Nucleus sampling probability mass |
|
||||
| `topK` | integer | 1–100 | Top-K token candidates per step |
|
||||
| `systemPrompt` | string | — | Global system prompt (null reverts to hardcoded default) |
|
||||
|
||||
Settings are persisted to `data/settings.json` and read on every request —
|
||||
changes take effect immediately without a service restart.
|
||||
|
||||
Reference in New Issue
Block a user