Package
+
+ {packageRows.map((row) => (
+
+
+
+ ))}
+ {row.label}
+{row.value}
+
+
+ Homepage
+ + {pkg.homepage} + ++ WeaveLLM — a local AI desktop runtime: managed llama.cpp backend, + visual workflow editor, model management and an OpenAI-compatible + proxy. Everything on this page is build-time package data or the + actual stack — nothing is simulated. +
+{row.label}
+{row.value}
+Homepage
+ + {pkg.homepage} + +{row.label}
+{row.value}
+{row.note}
++ This console is admin-only and stores no credentials. The auth gate + (WEAVELLM_AUTH) and provider keys live server-side; the console only + probes the real backend surface and never simulates data. +
++ Exercise POST /v1/chat/completions against the real wire. The canonical + route is /playground. +
++ Per-workflow run history straight from the backend logs — latest run up + front, full history on expand. No auto-refresh: a completed run is + immutable, so data refreshes on demand. +
++ Live runtime summary — backend endpoint, model registry state and recent + workflows, refreshed automatically from the /api surface. +
++ The managed llama-server registry — activation state, PID and port come + directly from the backend models API. +
+
+ POST /v1/chat/completions against the real wire — JSON for
+ non-streaming, OpenAI-wire SSE with a terminal [DONE] for
+ stream. Responses render verbatim; no derived metrics.
+
+ Backend health, the endpoints this console talks to, the auth gate and + a live WebSocket handshake probe — all against the real surface. +
++ Read-only: the auth gate is probed live, and server-owned configuration + is shown honestly. Nothing here mutates the server. +
++ Stored workflow definitions with run, log and delete controls — powered + by the backend workflow API. +
+{description}
+ {#if children} + {@render children()} + {/if} +Loading executions…
+ {/if} + {:else if rows.length === 0} ++ Counts cover the logs fetched for {totals.loaded} of {totals.total} workflows + loaded below — per-workflow history, not a global aggregate. +
+ + {#if error !== null} + + {/if} + +Loading runs…
+ {:else if cell.status === "error"} +{cell.error}
+ +No runs yet for this workflow.
+ {:else} ++ Authorization required — configure WEAVELLM_AUTH properly on the + backend to unlock this surface. +
+ {/if} + +Loading models…
+ {/if} + {:else if rows.length === 0} +No matching pages.
+ {:else} +Active: {activeModelIdsLabel(models)}
+ {/if} +{value}
+ + {#if hint !== ""} +{hint}
+ {/if} + + {#if children} ++ Suggestions — local registry ids and gateway/<workflow> names. + External adapter ids resolve server-side; the server answers unknown + models with a 404 envelope. +
++ Requests leave this console unauthenticated — the server owns auth + (WEAVELLM_AUTH) and credentials never live here. +
+ {#if isBusy()} + + {/if} + ++ The response renders verbatim here — no derived metrics. Press Send + to run the request. +
+ {/if} + + {#if status === STATUS.CANCELLED} ++ Request cancelled. Closing the connection makes the server abort the + upstream call (relay contract). +
+ {/if} + + {#if error !== null && status === STATUS.ERROR} ++ {error.status > 0 ? `${error.message} (HTTP ${error.status})` : error.message} +
+code: {error.code}
+ {#if errorHint() !== null} +{errorHint()}
+ {/if} + {#if error.errors !== undefined && error.errors !== null} +{highlightJson(JSON.stringify(error.errors, null, 2))}
+ {highlightJson(JSON.stringify(completion, null, 2))}
+ Final content (verbatim)
+{streamContent}
+ {rawSse}
+ Connecting to /v1/chat/completions…
+ {/if} +{healthCell.error}
+ + {/if} +{authCell.error}
+ + {/if} +API origin
+{apiOrigin}
+ WebSocket origin
+{wsOrigin}
+ + The endpoints this console talks to (client config). In dev, the Astro + proxy forwards them to the local backend. +
+{wsCard.value}
+{wsCard.announce}
+ {#if ws.phase === "error"} +{ws.error}
+ {/if} +{wsCard.hint}
+{probe.error}
+ + {/if} +{row.label}
+{row.value}
+ {SERVER_ENV_NOTE}
+{item}Loading workflow…
+ {:else} +{entry.description}
+ {/if} +{nameError()}
+ {/if} + + + + + {#if error !== null} +{error}
+ {/if} + {#if errorDetails.length > 0} +{errorDetails.join("\n")}
+ {/if}
+ Output
+ {#if outcome === ""} +The workflow produced no text.
+ {:else} +{outcome}
+ {/if}
+ {:else}
+ {#if error !== null}
+ {error}
+ {/if} + + + {/if} +{error}
+ +Loading logs…
+ {:else if rows.length === 0} +No runs yet. Run this workflow to see execution history here.
+ {:else} +Loading workflows…
+ {/if} + {:else if rows.length === 0} +{rows.length} workflow{rows.length === 1 ? "" : "s"}
+ +Delete workflow "{modalName}"? This cannot be undone.
+ {#if deleteError !== null} +{deleteError}
+ {/if} +