Skip to content

~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.

Add Requesty as an OpenAI-compatible LLM provider (#5052)

Thibault Jaigu98aebcd86cd4
diff --git a/README.md b/README.md
index 4586a7048c..bf0e8a240c 100644
--- a/README.md
+++ b/README.md
@@ -427,6 +427,7 @@ search(query="agentic research frameworks", engine="github")
- Anthropic Claude
- Google Gemini
- 100+ models via OpenRouter
+- 100+ models via Requesty
### Custom Endpoints
- **OpenAI-Compatible Endpoint** — any service speaking the OpenAI chat-completions API (vLLM, llama.cpp, gateways)
diff --git a/docs/CONFIGURATION.md b/docs/CONFIGURATION.md
index d399fda712..8edcfb953b 100644
--- a/docs/CONFIGURATION.md
+++ b/docs/CONFIGURATION.md
@@ -151,6 +151,7 @@ They are not available in the Web UI because they are needed to start the applic
| `llm.openai_endpoint.url` | `LDR_LLM_OPENAI_ENDPOINT_URL` | `https://openrouter.ai/api/v1` | URL of a custom OpenAI-compatible API endpoint. Use for llama.cpp server, vLLM, Ollama, OpenRouter, Groq, Together, or any service that implements the OpenAI API format. Example for local llama.cpp: http://localhost:8000/v1 | LLM |
| `llm.openrouter.api_key` | `LDR_LLM_OPENROUTER_API_KEY` | `null` | API key to use for the OpenRouter provider. | SEARCH |
| `llm.provider` | `LDR_LLM_PROVIDER` | `ollama` | The LLM service provider. Choose local providers (Ollama, LM Studio, LlamaCpp) for free on-device inference, or cloud APIs (OpenAI, Anthropic, Google, OpenRouter) for hosted models. | LLM |
+| `llm.requesty.api_key` | `LDR_LLM_REQUESTY_API_KEY` | `null` | API key to use for the Requesty provider. | SEARCH |
| `llm.require_local_endpoint` | `LDR_LLM_REQUIRE_LOCAL_ENDPOINT` | `false` | When enabled, the LLM provider's resolved endpoint must be on a local network (loopback, RFC1918, link-local). Cloud-only providers (OpenAI, Anthropic, Google, OpenRouter) are blocked. Independent of egress_scope. | LLM |
| `llm.supports_max_tokens` | `LDR_LLM_SUPPORTS_MAX_TOKENS` | `true` | Feature flag for max_tokens parameter. Disable if your LLM API returns errors when max_tokens is set. When disabled, the parameter is not passed to the API. | LLM |
| `llm.temperature` | `LDR_LLM_TEMPERATURE` | `0.7` | Controls randomness in model outputs. Lower values (0.0-0.3) produce deterministic, factual responses; higher values (0.7-1.0) produce more creative and varied outputs. | LLM |
diff --git a/docs/architecture.md b/docs/architecture.md
index 7d4d61d49f..c55b2617b0 100644
--- a/docs/architecture.md
+++ b/docs/architecture.md
@@ -202,6 +202,7 @@ mindmap
Anthropic
Google
OpenRouter
+ Requesty
Output
Markdown
PDF
@@ -228,6 +229,7 @@ mindmap
| Anthropic | Cloud | Claude 3 family |
| Google | Cloud | Gemini models |
| OpenRouter | Cloud | 100+ models via single API |
+| Requesty | Cloud | 100+ models via single API |
### Search Engines
@@ -386,13 +388,14 @@ class SearchEngineCustom(BaseSearchEngine):
#### LLM Provider Integration
-Supports 9 LLM providers with auto-discovery:
+Supports 10 LLM providers with auto-discovery:
- Ollama (local)
- LM Studio (local)
- OpenAI
- Anthropic
- Google Gemini
- OpenRouter (100+ models)
+- Requesty (100+ models)
- DeepSeek
- Mistral
- Groq
diff --git a/src/local_deep_research/defaults/default_settings.json b/src/local_deep_research/defaults/default_settings.json
index 27ee3819fb..6e0791ec33 100644
--- a/src/local_deep_research/defaults/default_settings.json
+++ b/src/local_deep_research/defaults/default_settings.json
@@ -681,6 +681,10 @@
{
"label": "OpenRouter API",
"value": "openrouter"
+ },
+ {
+ "label": "Requesty API",
+ "value": "requesty"
}
],
"step": null,
diff --git a/src/local_deep_research/defaults/llm_providers/requesty_settings.json b/src/local_deep_research/defaults/llm_providers/requesty_settings.json
new file mode 100644
index 0000000000..969b464ce5
--- /dev/null
+++ b/src/local_deep_research/defaults/llm_providers/requesty_settings.json
@@ -0,0 +1,16 @@
+{
+ "llm.requesty.api_key": {
+ "category": "llm_general",
+ "description": "API key to use for the Requesty provider.",
+ "editable": true,
+ "max_value": null,
+ "min_value": null,
+ "name": "Requesty API Key",
+ "options": null,
+ "step": null,
+ "type": "SEARCH",
+ "ui_element": "password",
+ "value": null,
+ "visible": true
+ }
+}
diff --git a/src/local_deep_research/llm/providers/implementations/requesty.py b/src/local_deep_research/llm/providers/implementations/requesty.py
new file mode 100644
index 0000000000..1f45353a8a
--- /dev/null
+++ b/src/local_deep_research/llm/providers/implementations/requesty.py
@@ -0,0 +1,31 @@
+"""Requesty LLM provider for Local Deep Research."""
+
+from ..base import Exposure
+from ..openai_base import OpenAICompatibleProvider
+
+
+class RequestyProvider(OpenAICompatibleProvider):
+ """Requesty provider using OpenAI-compatible endpoint.
+
+ Requesty is an LLM gateway that provides access to many different models
+ through a unified OpenAI-compatible API, using ``provider/model`` naming
+ (e.g. ``openai/gpt-4o-mini``), automatically supporting current and future
+ models without needing code updates.
+ """
+
+ provider_name = "Requesty"
+ api_key_setting = "llm.requesty.api_key"
+ default_base_url = "https://router.requesty.ai/v1"
+ default_model = "" # User must explicitly pick a model — no silent fallback
+
+ # Metadata for auto-discovery
+ provider_key = "REQUESTY"
+ company_name = "Requesty"
+ is_cloud = True
+ # Egress exposure (ADR-0007): cloud inference sink — data leaves the box.
+ egress_exposure = Exposure.EXPOSING
+
+ @classmethod
+ def requires_auth_for_models(cls):
+ """Requesty requires authentication for listing models."""
+ return True
diff --git a/src/local_deep_research/security/egress/policy.py b/src/local_deep_research/security/egress/policy.py
index 3cb49b72c9..404f9d2590 100644
--- a/src/local_deep_research/security/egress/policy.py
+++ b/src/local_deep_research/security/egress/policy.py
@@ -682,6 +682,7 @@ _CLOUD_LLM_PROVIDERS = frozenset(
"anthropic",
"google",
"openrouter",
+ "requesty",
"orcarouter",
"deepseek",
"xai",
diff --git a/tests/llm_providers/implementations/test_requesty_provider.py b/tests/llm_providers/implementations/test_requesty_provider.py
new file mode 100644
index 0000000000..432416da12
--- /dev/null
+++ b/tests/llm_providers/implementations/test_requesty_provider.py
@@ -0,0 +1,228 @@
+"""Tests for Requesty LLM provider."""
+
+import pytest
+from unittest.mock import Mock, patch
+
+from local_deep_research.llm.providers.implementations.requesty import (
+ RequestyProvider,
+)
+
+
+class TestRequestyProviderMetadata:
+ """Tests for RequestyProvider class metadata."""
+
+ def test_provider_name(self):
+ """Provider name is correct."""
+ assert RequestyProvider.provider_name == "Requesty"
+
+ def test_provider_key(self):
+ """Provider key is correct."""
+ assert RequestyProvider.provider_key == "REQUESTY"
+
+ def test_is_cloud(self):
+ """Requesty is a cloud provider."""
+ assert RequestyProvider.is_cloud is True
+
+ def test_company_name(self):
+ """Company name is Requesty."""
+ assert RequestyProvider.company_name == "Requesty"
+
+ def test_api_key_setting(self):
+ """API key setting is correct."""
+ assert RequestyProvider.api_key_setting == "llm.requesty.api_key"
+
+ def test_default_model(self):
+ """Default model is empty by design — users must explicitly pick one."""
+ assert RequestyProvider.default_model == ""
+
+ def test_default_base_url(self):
+ """Default base URL is correct."""
+ assert "router.requesty.ai" in RequestyProvider.default_base_url
+
+
+class TestRequestyCreateLLM:
+ """Tests for create_llm method."""
+
+ def test_create_llm_raises_without_api_key(self):
+ """Raises ValueError when API key not configured."""
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.return_value = None
+
+ with pytest.raises(ValueError) as exc_info:
+ RequestyProvider.create_llm()
+
+ assert "api key" in str(exc_info.value).lower()
+
+ def test_create_llm_with_valid_api_key(self):
+ """Successfully creates ChatOpenAI instance with valid API key."""
+
+ def mock_get_setting_side_effect(key, default=None, *args, **kwargs):
+ settings_map = {
+ "llm.requesty.api_key": "test-requesty-key",
+ "llm.max_tokens": None,
+ "llm.streaming": None,
+ "llm.max_retries": None,
+ "llm.request_timeout": None,
+ }
+ return settings_map.get(key, default)
+
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.side_effect = mock_get_setting_side_effect
+
+ with patch(
+ "local_deep_research.llm.providers.openai_base.ChatOpenAI"
+ ) as mock_chat:
+ mock_llm = Mock()
+ mock_chat.return_value = mock_llm
+
+ result = RequestyProvider.create_llm(model_name="test-model")
+
+ assert result is mock_llm
+ mock_chat.assert_called_once()
+
+ def test_create_llm_uses_default_model_when_none(self):
+ """Raises ValueError when no model name is provided (no silent default)."""
+
+ def mock_get_setting_side_effect(key, default=None, *args, **kwargs):
+ settings_map = {
+ "llm.requesty.api_key": "test-key",
+ "llm.max_tokens": None,
+ "llm.streaming": None,
+ "llm.max_retries": None,
+ "llm.request_timeout": None,
+ }
+ return settings_map.get(key, default)
+
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.side_effect = mock_get_setting_side_effect
+
+ with pytest.raises(ValueError, match="model not configured"):
+ RequestyProvider.create_llm()
+
+ def test_create_llm_with_custom_model(self):
+ """Uses custom model when specified (provider/model naming)."""
+
+ def mock_get_setting_side_effect(key, default=None, *args, **kwargs):
+ settings_map = {
+ "llm.requesty.api_key": "test-key",
+ "llm.max_tokens": None,
+ "llm.streaming": None,
+ "llm.max_retries": None,
+ "llm.request_timeout": None,
+ }
+ return settings_map.get(key, default)
+
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.side_effect = mock_get_setting_side_effect
+
+ with patch(
+ "local_deep_research.llm.providers.openai_base.ChatOpenAI"
+ ) as mock_chat:
+ RequestyProvider.create_llm(model_name="openai/gpt-4o-mini")
+
+ call_kwargs = mock_chat.call_args[1]
+ assert call_kwargs["model"] == "openai/gpt-4o-mini"
+
+ def test_create_llm_passes_temperature(self):
+ """Passes temperature parameter."""
+
+ def mock_get_setting_side_effect(key, default=None, *args, **kwargs):
+ settings_map = {
+ "llm.requesty.api_key": "test-key",
+ "llm.max_tokens": None,
+ "llm.streaming": None,
+ "llm.max_retries": None,
+ "llm.request_timeout": None,
+ }
+ return settings_map.get(key, default)
+
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.side_effect = mock_get_setting_side_effect
+
+ with patch(
+ "local_deep_research.llm.providers.openai_base.ChatOpenAI"
+ ) as mock_chat:
+ RequestyProvider.create_llm(
+ model_name="test-model", temperature=0.5
+ )
+
+ call_kwargs = mock_chat.call_args[1]
+ assert call_kwargs["temperature"] == 0.5
+
+ def test_create_llm_uses_requesty_base_url(self):
+ """Uses Requesty's base URL."""
+
+ def mock_get_setting_side_effect(key, default=None, *args, **kwargs):
+ settings_map = {
+ "llm.requesty.api_key": "test-key",
+ "llm.max_tokens": None,
+ "llm.streaming": None,
+ "llm.max_retries": None,
+ "llm.request_timeout": None,
+ }
+ return settings_map.get(key, default)
+
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.side_effect = mock_get_setting_side_effect
+
+ with patch(
+ "local_deep_research.llm.providers.openai_base.ChatOpenAI"
+ ) as mock_chat:
+ RequestyProvider.create_llm(model_name="test-model")
+
+ call_kwargs = mock_chat.call_args[1]
+ assert "router.requesty.ai" in call_kwargs["base_url"]
+
+
+class TestRequestyIsAvailable:
+ """Tests for is_available method."""
+
+ def test_is_available_true_when_key_exists(self):
+ """Returns True when API key is configured."""
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.return_value = "test-key"
+
+ result = RequestyProvider.is_available()
+ assert result is True
+
+ def test_is_available_false_when_no_key(self):
+ """Returns False when API key is not configured."""
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.return_value = None
+
+ result = RequestyProvider.is_available()
+ assert result is False
+
+ def test_is_available_false_when_empty_key(self):
+ """Returns False when API key is empty string."""
+ with patch(
+ "local_deep_research.config.thread_settings.get_setting_from_snapshot"
+ ) as mock_get_setting:
+ mock_get_setting.return_value = ""
+
+ result = RequestyProvider.is_available()
+ assert result is False
+
+
+class TestRequestyRequiresAuth:
+ """Tests for requires_auth_for_models method."""
+
+ def test_requires_auth_for_models(self):
+ """Requesty requires authentication for listing models."""
+ assert RequestyProvider.requires_auth_for_models() is True
diff --git a/tests/settings/golden_master_settings.json b/tests/settings/golden_master_settings.json
index 80adf888e5..979d162434 100644
--- a/tests/settings/golden_master_settings.json
+++ b/tests/settings/golden_master_settings.json
@@ -1514,6 +1514,10 @@
{
"label": "OpenRouter API",
"value": "openrouter"
+ },
+ {
+ "label": "Requesty API",
+ "value": "requesty"
}
],
"step": null,
@@ -1522,6 +1526,20 @@
"value": "ollama",
"visible": true
},
+ "llm.requesty.api_key": {
+ "category": "llm_general",
+ "description": "API key to use for the Requesty provider.",
+ "editable": true,
+ "max_value": null,
+ "min_value": null,
+ "name": "Requesty API Key",
+ "options": null,
+ "step": null,
+ "type": "SEARCH",
+ "ui_element": "password",
+ "value": null,
+ "visible": true
+ },
"llm.require_local_endpoint": {
"category": "policy",
"description": "When enabled, the LLM provider's resolved endpoint must be on a local network (loopback, RFC1918, link-local). Cloud-only providers (OpenAI, Anthropic, Google, OpenRouter) are blocked. Independent of egress_scope.",
diff --git a/tests/settings/test_settings_defaults_integrity.py b/tests/settings/test_settings_defaults_integrity.py
index bbdea993d6..112c1485f8 100644
--- a/tests/settings/test_settings_defaults_integrity.py
+++ b/tests/settings/test_settings_defaults_integrity.py
@@ -222,6 +222,7 @@ def _is_dynamic_setting(key: str) -> bool:
"llm.google.",
"llm.ionos.",
"llm.openrouter.",
+ "llm.requesty.",
"llm.orcarouter.",
"llm.xai.",
"llm.ollama.",