fix(embed): allow separate OPENAI_EMBEDDING_BASE_URL + OPENAI_EMBEDDING_API_KEY

The OpenAI-compat embedding provider previously read OPENAI_BASE_URL and
OPENAI_API_KEY only, which the chat-LLM path (src/providers/openai.ts)
also reads. That couples both calls to the same endpoint — so operators
who want, say, chat completions on a fast hosted provider (Novita /
DeepInfra) and embeddings on a self-hosted vLLM cluster (Qwen3-Embedding-8B
on a dedicated GPU) had to either move both to the same endpoint or run
agentmemory against a single provider with whatever embedding model it
happens to expose.

Add two embedding-scoped overrides with fallback to the existing vars:

  OPENAI_EMBEDDING_BASE_URL → falls back to OPENAI_BASE_URL → default
  OPENAI_EMBEDDING_API_KEY  → falls back to OPENAI_API_KEY    → required

The fallback chain keeps existing setups working without any .env
changes. New setups can mix and match — common patterns:

  # vLLM (self-hosted GPU, free, batchable) for embeddings + Novita (DeepSeek V4 Flash) for chat
  OPENAI_BASE_URL=https://api.novita.ai/v3/openai
  OPENAI_API_KEY=sk-novita-...
  OPENAI_EMBEDDING_BASE_URL=https://embed.your.lan
  OPENAI_EMBEDDING_API_KEY=local-no-auth    # endpoints that ignore Authorization
  OPENAI_EMBEDDING_MODEL=Qwen3-Embedding-8B
  OPENAI_EMBEDDING_DIMENSIONS=4096

  # Local Ollama for embeddings + remote for chat
  OPENAI_BASE_URL=https://api.openai.com
  OPENAI_API_KEY=sk-...
  OPENAI_EMBEDDING_BASE_URL=http://localhost:11434
  OPENAI_EMBEDDING_API_KEY=ollama
  OPENAI_EMBEDDING_MODEL=nomic-embed-text

The separate API key matters because most local endpoints
(Ollama / LM Studio / llama.cpp / vLLM) ignore Authorization entirely
but Node fetch still requires a non-empty Bearer token. Setting
OPENAI_EMBEDDING_API_KEY=anything-truthy unblocks that case without
revealing the real OPENAI_API_KEY to whatever's on localhost.

No code-paths other than the embedding provider are touched. Reviewed
against the 17 existing test cases in test/embedding-provider.test.ts;
no regression (the 6 pre-existing failures on main are env-pollution
when ~/.agentmemory/.env has API keys set, unrelated to this change).

Author: efenex

src/providers/embedding/openai.ts

@@ -48,16 +48,30 @@ function resolveDimensions(model: string, override: string | undefined): number
  * `api-key` header instead of `Authorization: Bearer`.
  *
  * Required env vars:
- *   OPENAI_API_KEY            — API key
+ *   OPENAI_API_KEY               — API key (fallback for OPENAI_EMBEDDING_API_KEY)
  *
  * Optional:
- *   OPENAI_BASE_URL           — base URL without path (default: https://api.openai.com).
- *                               Azure: https://<resource>.openai.azure.com/openai/deployments/<deployment>
- *   OPENAI_API_VERSION        — Azure api-version query param (default: 2024-08-01-preview)
- *   OPENAI_EMBEDDING_MODEL    — model name (default: text-embedding-3-small)
- *   OPENAI_EMBEDDING_DIMENSIONS — override reported dimensions (required for
- *                                 custom / self-hosted models not in the
- *                                 MODEL_DIMENSIONS table above)
+ *   OPENAI_BASE_URL              — base URL without path (default: https://api.openai.com).
+ *                                  Azure: https://<resource>.openai.azure.com/openai/deployments/<deployment>
+ *   OPENAI_EMBEDDING_BASE_URL    — embedding-specific base URL override (defaults
+ *                                  to OPENAI_BASE_URL). Lets operators run
+ *                                  embeddings on a separate endpoint from chat —
+ *                                  e.g. local Ollama / LM Studio / llama.cpp /
+ *                                  vLLM at http://localhost:1234 for unlimited
+ *                                  free embeddings, while keeping chat
+ *                                  completions on a rate-limited but high-quality
+ *                                  hosted provider. Azure detection runs on
+ *                                  whichever URL ends up selected.
+ *   OPENAI_EMBEDDING_API_KEY     — separate API key for the embedding endpoint
+ *                                  (defaults to OPENAI_API_KEY). Useful when the
+ *                                  embedding endpoint requires a different key
+ *                                  or no key at all (set to e.g. "local" for
+ *                                  endpoints that ignore Authorization).
+ *   OPENAI_API_VERSION           — Azure api-version query param (default: 2024-08-01-preview)
+ *   OPENAI_EMBEDDING_MODEL       — model name (default: text-embedding-3-small)
+ *   OPENAI_EMBEDDING_DIMENSIONS  — override reported dimensions (required for
+ *                                  custom / self-hosted models not in the
+ *                                  MODEL_DIMENSIONS table above)
  */
 export class OpenAIEmbeddingProvider implements EmbeddingProvider {
   readonly name = "openai";
@@ -69,9 +83,26 @@ export class OpenAIEmbeddingProvider implements EmbeddingProvider {
   private azureApiVersion: string;
 
   constructor(apiKey?: string) {
-    this.apiKey = apiKey || getEnvVar("OPENAI_API_KEY") || "";
-    if (!this.apiKey) throw new Error("OPENAI_API_KEY is required");
-    this.baseUrl = normalizeBaseUrl(getEnvVar("OPENAI_BASE_URL"));
+    // Separate API key path: caller-passed wins, then OPENAI_EMBEDDING_API_KEY,
+    // then fall back to OPENAI_API_KEY. Allows e.g. a placeholder key for
+    // local endpoints that ignore Authorization (most do).
+    this.apiKey =
+      apiKey ||
+      getEnvVar("OPENAI_EMBEDDING_API_KEY") ||
+      getEnvVar("OPENAI_API_KEY") ||
+      "";
+    if (!this.apiKey) {
+      throw new Error(
+        "API key is required (via constructor, OPENAI_EMBEDDING_API_KEY, or OPENAI_API_KEY)",
+      );
+    }
+    // Embedding-specific base URL override; falls back to OPENAI_BASE_URL,
+    // then normalizeBaseUrl's default. The chat-LLM path (src/providers/openai.ts)
+    // still reads only OPENAI_BASE_URL, so setting OPENAI_EMBEDDING_BASE_URL
+    // alone moves embeddings to the new endpoint without affecting chat.
+    this.baseUrl = normalizeBaseUrl(
+      getEnvVar("OPENAI_EMBEDDING_BASE_URL") || getEnvVar("OPENAI_BASE_URL"),
+    );
     this.model = getEnvVar("OPENAI_EMBEDDING_MODEL") || DEFAULT_MODEL;
     this.dimensions = resolveDimensions(
       this.model,

test/embedding-provider.test.ts

@@ -74,7 +74,8 @@ describe("OpenAIEmbeddingProvider", () => {
 
   it("throws when no API key is provided", () => {
     delete process.env["OPENAI_API_KEY"];
-    expect(() => new OpenAIEmbeddingProvider()).toThrow("OPENAI_API_KEY is required");
+    delete process.env["OPENAI_EMBEDDING_API_KEY"];
+    expect(() => new OpenAIEmbeddingProvider()).toThrow(/API key is required.*OPENAI_EMBEDDING_API_KEY.*OPENAI_API_KEY/);
   });
 
   it("respects OPENAI_BASE_URL env var", async () => {