feat: add Requesty as an OpenAI-compatible provider (#2399)

Requesty is an OpenAI-compatible LLM gateway. This mirrors the existing
OpenRouter named-provider wiring 1:1:

- llm_wrapper.py / openai_compatible_llm.py: add "requesty" to the
  provider lists and a base_url branch -> https://router.requesty.ai/v1
- config.py: default model map (openai/gpt-4o-mini), embeddings env vars,
  dataclass fields, and from_env wiring (REQUESTY_API_KEY fallbacks)
- embeddings.py: requesty branch (same /v1 base) + Supported list
- docs/llmProviders.json: factual provider entries

Tested live against https://router.requesty.ai/v1/chat/completions
(model openai/gpt-4o-mini) -> HTTP 200.

Author: Thibaultjaigu

hindsight-api-slim/hindsight_api/config.py

@@ -282,6 +282,11 @@ def parse_gemini_service_tier(value: str | None) -> str | None:
 ENV_RERANKER_OPENROUTER_MODEL = "HINDSIGHT_API_RERANKER_OPENROUTER_MODEL"
 ENV_RERANKER_OPENROUTER_BASE_URL = "HINDSIGHT_API_RERANKER_OPENROUTER_BASE_URL"
 
+# Requesty configuration (OpenAI-compatible gateway; embeddings)
+ENV_REQUESTY_API_KEY = "HINDSIGHT_API_REQUESTY_API_KEY"
+ENV_EMBEDDINGS_REQUESTY_API_KEY = "HINDSIGHT_API_EMBEDDINGS_REQUESTY_API_KEY"
+ENV_EMBEDDINGS_REQUESTY_MODEL = "HINDSIGHT_API_EMBEDDINGS_REQUESTY_MODEL"
+
 # ZeroEntropy configuration (embeddings)
 ENV_EMBEDDINGS_ZEROENTROPY_API_KEY = "HINDSIGHT_API_EMBEDDINGS_ZEROENTROPY_API_KEY"
 ENV_EMBEDDINGS_ZEROENTROPY_MODEL = "HINDSIGHT_API_EMBEDDINGS_ZEROENTROPY_MODEL"
@@ -643,6 +648,7 @@ def parse_gemini_service_tier(value: str | None) -> str | None:
     "bedrock": "us.amazon.nova-2-lite-v1:0",
     "volcano": "doubao-pro-32k",
     "openrouter": "qwen/qwen3.5-9b",
+    "requesty": "openai/gpt-4o-mini",
     "fireworks": "accounts/fireworks/models/llama-v3p1-8b-instruct",
     "nous": "deepseek/deepseek-v4-flash",
 }
@@ -799,6 +805,9 @@ def _parse_strategy_boosts(raw: str | None) -> dict[str, str]:
 DEFAULT_RERANKER_OPENROUTER_MODEL = "cohere/rerank-v3.5"
 DEFAULT_RERANKER_OPENROUTER_BASE_URL = "https://openrouter.ai/api/v1/rerank"
 
+# Requesty defaults
+DEFAULT_EMBEDDINGS_REQUESTY_MODEL = "openai/text-embedding-3-small"
+
 # ZeroEntropy defaults
 DEFAULT_EMBEDDINGS_ZEROENTROPY_MODEL = "zembed-1"
 # Shared between embeddings (zembed-1) and reranker (zerank-*) — the host is the same.
@@ -1600,6 +1609,8 @@ class HindsightConfig:
     embeddings_cohere_output_dimensions: int | None
     embeddings_openrouter_api_key: str | None
     embeddings_openrouter_model: str
+    embeddings_requesty_api_key: str | None
+    embeddings_requesty_model: str
     embeddings_litellm_api_base: str
     embeddings_litellm_api_key: str | None
     embeddings_litellm_model: str
@@ -2415,6 +2426,11 @@ def from_env(cls) -> "HindsightConfig":
             or os.getenv(ENV_OPENROUTER_API_KEY)
             or os.getenv(ENV_LLM_API_KEY),
             embeddings_openrouter_model=os.getenv(ENV_EMBEDDINGS_OPENROUTER_MODEL, DEFAULT_EMBEDDINGS_OPENROUTER_MODEL),
+            # Requesty embeddings (with fallback to shared Requesty key, then LLM key)
+            embeddings_requesty_api_key=os.getenv(ENV_EMBEDDINGS_REQUESTY_API_KEY)
+            or os.getenv(ENV_REQUESTY_API_KEY)
+            or os.getenv(ENV_LLM_API_KEY),
+            embeddings_requesty_model=os.getenv(ENV_EMBEDDINGS_REQUESTY_MODEL, DEFAULT_EMBEDDINGS_REQUESTY_MODEL),
             # ZeroEntropy embeddings
             embeddings_zeroentropy_api_key=os.getenv(ENV_EMBEDDINGS_ZEROENTROPY_API_KEY)
             or os.getenv("ZEROENTROPY_API_KEY"),

hindsight-api-slim/hindsight_api/engine/embeddings.py

@@ -1638,6 +1638,20 @@ def create_embeddings_from_env() -> Embeddings:
             batch_size=config.embeddings_openai_batch_size,
             dimensions=config.embeddings_openai_dimensions,
         )
+    elif provider == "requesty":
+        api_key = config.embeddings_requesty_api_key
+        if not api_key:
+            raise ValueError(
+                "HINDSIGHT_API_EMBEDDINGS_REQUESTY_API_KEY, HINDSIGHT_API_REQUESTY_API_KEY, "
+                f"or {ENV_LLM_API_KEY} is required when {ENV_EMBEDDINGS_PROVIDER} is 'requesty'"
+            )
+        return OpenAIEmbeddings(
+            api_key=api_key,
+            model=config.embeddings_requesty_model,
+            base_url="https://router.requesty.ai/v1",
+            batch_size=config.embeddings_openai_batch_size,
+            dimensions=config.embeddings_openai_dimensions,
+        )
     elif provider == "zeroentropy":
         api_key = config.embeddings_zeroentropy_api_key
         if not api_key:
@@ -1701,6 +1715,6 @@ def create_embeddings_from_env() -> Embeddings:
     else:
         raise ValueError(
             f"Unknown embeddings provider: {provider}. "
-            f"Supported: 'local', 'onnx', 'tei', 'openai', 'openai-codex', 'openrouter', 'cohere', 'google', "
+            f"Supported: 'local', 'onnx', 'tei', 'openai', 'openai-codex', 'openrouter', 'requesty', 'cohere', 'google', "
             f"'zeroentropy', 'litellm', 'litellm-sdk'"
         )

hindsight-api-slim/hindsight_api/engine/llm_wrapper.py

@@ -464,6 +464,7 @@ def create_llm_provider(
         "deepseek",
         "volcano",
         "openrouter",
+        "requesty",
         "zai",
         "opencode-go",
         "atlas",
@@ -592,6 +593,7 @@ def __init__(
             "bedrock",
             "volcano",
             "openrouter",
+            "requesty",
             "zai",
             "opencode-go",
             "atlas",
@@ -617,6 +619,8 @@ def __init__(
                 self.base_url = "https://api.deepseek.com"
             elif self.provider == "openrouter":
                 self.base_url = "https://openrouter.ai/api/v1"
+            elif self.provider == "requesty":
+                self.base_url = "https://router.requesty.ai/v1"
             elif self.provider == "zai":
                 self.base_url = "https://api.z.ai/api/coding/paas/v4"
             elif self.provider == "opencode-go":

hindsight-api-slim/hindsight_api/engine/providers/openai_compatible_llm.py

@@ -465,6 +465,7 @@ def __init__(
             "deepseek",
             "volcano",
             "openrouter",
+            "requesty",
             "zai",
             "opencode-go",
             "atlas",
@@ -489,6 +490,8 @@ def __init__(
                 self.base_url = "https://api.deepseek.com"
             elif self.provider == "openrouter":
                 self.base_url = "https://openrouter.ai/api/v1"
+            elif self.provider == "requesty":
+                self.base_url = "https://router.requesty.ai/v1"
             elif self.provider == "zai":
                 self.base_url = "https://api.z.ai/api/coding/paas/v4"
             elif self.provider == "opencode-go":
@@ -513,6 +516,7 @@ def __init__(
                 "minimax",
                 "deepseek",
                 "openrouter",
+                "requesty",
                 "zai",
                 "opencode-go",
                 "atlas",

hindsight-docs/docs/developer/configuration.md

@@ -164,7 +164,7 @@ For non-English banks (especially CJK) and the language/extraction-language trad
 
 | Variable | Description | Default |
 |----------|-------------|---------|
-| `HINDSIGHT_API_LLM_PROVIDER` | Provider: `openai`, `openai-codex`, `claude-code`, `anthropic`, `gemini`, `groq`, `minimax`, `deepseek`, `zai`, `opencode-go`, `nous`, `fireworks`, `ollama`, `ollama-cloud`, `lmstudio`, `llamacpp`, `vertexai`, `bedrock`, `litellm`, `litellmrouter`, `volcano`, `openrouter`, `none` | `openai` |
+| `HINDSIGHT_API_LLM_PROVIDER` | Provider: `openai`, `openai-codex`, `claude-code`, `anthropic`, `gemini`, `groq`, `minimax`, `deepseek`, `zai`, `opencode-go`, `nous`, `fireworks`, `ollama`, `ollama-cloud`, `lmstudio`, `llamacpp`, `vertexai`, `bedrock`, `litellm`, `litellmrouter`, `volcano`, `openrouter`, `requesty`, `none` | `openai` |
 | `HINDSIGHT_API_LLM_API_KEY` | API key for LLM provider | - |
 | `HINDSIGHT_API_LLM_MODEL` | Model name | `gpt-5-mini` |
 | `HINDSIGHT_API_LLM_BASE_URL` | Custom LLM endpoint | Provider default |
@@ -270,6 +270,11 @@ export HINDSIGHT_API_LLM_PROVIDER=openrouter
 export HINDSIGHT_API_LLM_API_KEY=your-openrouter-api-key
 export HINDSIGHT_API_LLM_MODEL=qwen/qwen3.5-9b
 
+# Requesty (OpenAI-compatible gateway)
+export HINDSIGHT_API_LLM_PROVIDER=requesty
+export HINDSIGHT_API_LLM_API_KEY=your-requesty-api-key
+export HINDSIGHT_API_LLM_MODEL=openai/gpt-4o-mini
+
 # DeepSeek (OpenAI-compatible, https://api.deepseek.com)
 export HINDSIGHT_API_LLM_PROVIDER=deepseek
 export HINDSIGHT_API_LLM_API_KEY=sk-xxxxxxxxxxxx
@@ -518,7 +523,7 @@ two slots that retain/consolidation cannot consume.
 
 | Variable | Description | Default |
 |----------|-------------|---------|
-| `HINDSIGHT_API_EMBEDDINGS_PROVIDER` | Provider: `local`, `onnx`, `tei`, `openai`, `openai-codex`, `openrouter`, `cohere`, `google`, `zeroentropy`, `litellm`, or `litellm-sdk` | `local` |
+| `HINDSIGHT_API_EMBEDDINGS_PROVIDER` | Provider: `local`, `onnx`, `tei`, `openai`, `openai-codex`, `openrouter`, `requesty`, `cohere`, `google`, `zeroentropy`, `litellm`, or `litellm-sdk` | `local` |
 | `HINDSIGHT_API_EMBEDDINGS_LOCAL_MODEL` | Model for local provider | `BAAI/bge-small-en-v1.5` |
 | `HINDSIGHT_API_EMBEDDINGS_LOCAL_TRUST_REMOTE_CODE` | Allow loading models with custom code (security risk, disabled by default) | `false` |
 | `HINDSIGHT_API_EMBEDDINGS_LOCAL_FORCE_CPU` | Force CPU mode for local embeddings (avoids MPS/XPC issues on macOS) | `false` |
@@ -540,6 +545,8 @@ two slots that retain/consolidation cannot consume.
 | `HINDSIGHT_API_EMBEDDINGS_OPENAI_BATCH_SIZE` | Max inputs per `embeddings.create` call for `openai`/`openrouter` providers — lower this when the upstream endpoint enforces stricter limits (e.g. DashScope caps at 10) | `100` |
 | `HINDSIGHT_API_EMBEDDINGS_OPENAI_DIMENSIONS` | Optional requested output dimensions for OpenAI `text-embedding-3` models (e.g., `384` to match an existing pgvector schema) | - |
 | `HINDSIGHT_API_EMBEDDINGS_OPENROUTER_API_KEY` | OpenRouter API key for embeddings (falls back to `HINDSIGHT_API_OPENROUTER_API_KEY`, then `HINDSIGHT_API_LLM_API_KEY`) | - |
+| `HINDSIGHT_API_EMBEDDINGS_REQUESTY_API_KEY` | Requesty API key for embeddings (falls back to `HINDSIGHT_API_REQUESTY_API_KEY`, then `HINDSIGHT_API_LLM_API_KEY`) | - |
+| `HINDSIGHT_API_EMBEDDINGS_REQUESTY_MODEL` | Requesty embedding model | `openai/text-embedding-3-small` |
 | `HINDSIGHT_API_EMBEDDINGS_OPENROUTER_MODEL` | OpenRouter embedding model | `perplexity/pplx-embed-v1-0.6b` |
 | `HINDSIGHT_API_EMBEDDINGS_ZEROENTROPY_API_KEY` | ZeroEntropy API key for embeddings | - |
 | `HINDSIGHT_API_EMBEDDINGS_ZEROENTROPY_MODEL` | ZeroEntropy embedding model | `zembed-1` |

hindsight-docs/src/data/llmProviders.json

@@ -15,6 +15,7 @@
   {"id": "atlas",         "label": "Atlas Cloud",       "iconKey": "openai-compatible",  "defaultModel": "deepseek-ai/deepseek-v4-pro"},
   {"id": "volcano",       "label": "Volcano Engine",    "iconKey": "zap",                "defaultModel": "doubao-pro-32k"},
   {"id": "openrouter",    "label": "OpenRouter",        "iconKey": "globe",              "defaultModel": "qwen/qwen3.5-9b"},
+  {"id": "requesty",      "label": "Requesty",          "iconKey": "openai-compatible",  "defaultModel": "openai/gpt-4o-mini"},
   {"id": "openai-codex",  "label": "OpenAI Codex",      "iconKey": "openai",             "defaultModel": "gpt-5.4-mini"},
   {"id": "claude-code",   "label": "Claude Code",       "iconKey": "anthropic",          "defaultModel": "claude-sonnet-4-5-20250929"},
   {"id": "bedrock",       "label": "AWS Bedrock",       "iconKey": "cloud",              "defaultModel": "us.amazon.nova-2-lite-v1:0"},