feat(connectors): add Ollama and LM Studio as a local auth kind

Both servers expose OpenAI-compatible /v1/chat/completions and /v1/models,
so we route them through pi-ai's `openai-completions` API the same way
Fireworks does — but with a per-installation base URL the user enters
during onboarding instead of a hardcoded one. The shared schemas grow a
`local` credential payload carrying just the baseUrl; the credential
resolver surfaces it on `ResolvedCredential.baseUrl`, the runtime bridge
forwards it through `prepareRun`, and the worker plumbs it into
`resolveModel` so the streamed Model<openai-completions> targets the
user's server. The local-cost path short-circuits to zero in cost-writer
since users bill their own hardware, and the live model list is fetched
from the saved baseUrl by `fetchLocalConnectorModels` (Ollama and LM
Studio both speak the standard OpenAI catalog shape). Onboarding ships
a new LocalProviderCard and the Settings panel a LocalForm; both default
to the catalog's per-connector localhost URL and accept any custom URL.

src/main/agent-runtime/cost-writer.ts

@@ -1,4 +1,4 @@
-import { calculateCost, type Usage, type Api, type Model } from "@mariozechner/pi-ai";
+import { calculateCost, type Usage } from "@mariozechner/pi-ai";
 import type { AppDatabase } from "@main/database";
 import { getConnectorDefinition } from "@main/connectors";
 import { resolveModel, resolveModelCostFallback } from "@main/model-resolver";
@@ -17,6 +17,12 @@ interface ObserveInput {
   usage: Usage;
 }
 
+// Local OpenAI-compatible servers run on the user's own machine, so the
+// LLM call has no third-party billing line. Record a zero cost entry
+// without invoking pi-ai's calculator (which would require the model's
+// baseUrl from the credential — unnecessary work for a flat-zero cost).
+const ZERO_COST_PROVIDERS: ReadonlySet<string> = new Set(["ollama", "lm-studio"]);
+
 export class CostWriter {
   constructor(private readonly deps: CostWriterDeps) {}
 
@@ -25,6 +31,18 @@ export class CostWriter {
     if (!definition.provider) {
       throw new Error(`[cost-writer] connector ${input.connectorId} has no provider mapping`);
     }
+    if (ZERO_COST_PROVIDERS.has(definition.provider)) {
+      this.deps.db.core.runs.recordCostEntry({
+        runId: input.runId,
+        companyId: input.companyId,
+        agentId: input.agentId,
+        taskId: input.taskId,
+        connectorId: input.connectorId,
+        amountUsd: 0,
+        unattributed: false,
+      });
+      return;
+    }
     const model = resolveModel(definition.provider, input.modelId);
     if (!model) {
       throw new Error(`[cost-writer] model ${input.modelId} not found for provider ${definition.provider}`);

src/main/agent-runtime/runtime-bridge.ts

@@ -440,6 +440,7 @@ export class RuntimeBridge {
         provider: definition.provider,
         apiKey: resolved.apiKey,
         extraEnv: resolved.extraEnv ?? {},
+        baseUrl: resolved.baseUrl ?? null,
         priorMessages,
       },
     };