Merge branch 'v0.3.0' into claude-code

# Conflicts:
#	routstr/upstream/base.py

routstr/core/main.py

@@ -66,11 +66,6 @@ async def lifespan(_: FastAPI) -> AsyncGenerator[None, None]:
 
                 await reset_all_reserved_balances(session)
 
-        if not s.admin_password:
-            logger.warning(
-                f"Admin password is not set. Visit {s.http_url or 'http://localhost:8000'}/admin to set the password."
-            )
-
         # Apply app metadata from settings
         try:
             app.title = s.name

routstr/core/settings.py

@@ -6,7 +6,7 @@ import os
 from datetime import datetime, timezone
 from typing import Any
 
-from pydantic.v1 import BaseModel, BaseSettings, Field
+from pydantic.v1 import BaseModel, BaseSettings, Field, validator
 from sqlmodel.ext.asyncio.session import AsyncSession
 
 
@@ -37,6 +37,13 @@ class Settings(BaseSettings):
 
     # Cashu
     cashu_mints: list[str] = Field(default_factory=list, env="CASHU_MINTS")
+
+    @validator("cashu_mints", pre=True, each_item=True)
+    def normalize_mint_url(cls, v: str) -> str:
+        if isinstance(v, str):
+            return v.rstrip("/")
+        return v
+
     receive_ln_address: str = Field(default="", env="RECEIVE_LN_ADDRESS")
     primary_mint: str = Field(default="", env="PRIMARY_MINT_URL")
     primary_mint_unit: str = Field(default="sat", env="PRIMARY_MINT_UNIT")
@@ -62,7 +69,7 @@ class Settings(BaseSettings):
     cors_origins: list[str] = Field(default_factory=lambda: ["*"], env="CORS_ORIGINS")
     tor_proxy_url: str = Field(default="socks5://127.0.0.1:9050", env="TOR_PROXY_URL")
     providers_refresh_interval_seconds: int = Field(
-        default=0, env="PROVIDERS_REFRESH_INTERVAL_SECONDS"
+        default=300, env="PROVIDERS_REFRESH_INTERVAL_SECONDS"
     )
     pricing_refresh_interval_seconds: int = Field(
         default=120, env="PRICING_REFRESH_INTERVAL_SECONDS"

routstr/discovery.py

@@ -6,7 +6,7 @@ from typing import Any
 
 import httpx
 import websockets
-from fastapi import APIRouter, HTTPException
+from fastapi import APIRouter
 
 from .core.logging import get_logger
 from .core.settings import settings
@@ -389,9 +389,6 @@ async def get_providers(
     Return cached providers. If include_json, return provider+health; otherwise provider only.
     Optional filter by pubkey.
     """
-    if settings.providers_refresh_interval_seconds == 0:
-        raise HTTPException(status_code=404, detail="Provider discovery is disabled")
-
     cache = await get_cache()
     if not cache:
         await refresh_providers_cache(pubkey=pubkey)

routstr/payment/models.py

@@ -253,11 +253,6 @@ async def list_models(
             else 1.01,
         )
         for r in rows
-        if include_disabled
-        or (
-            r.upstream_provider_id in providers_by_id
-            and providers_by_id[r.upstream_provider_id].enabled
-        )
     ]
 
 
@@ -270,8 +265,6 @@ async def get_model_by_id(
     if not row or not row.enabled:
         return None
     provider = await session.get(UpstreamProviderRow, provider_id)
-    if not provider or not provider.enabled:
-        return None
     provider_fee = provider.provider_fee if provider else 1.01
     return _row_to_model(row, apply_provider_fee=True, provider_fee=provider_fee)
 

routstr/payment/price.py

@@ -79,29 +79,15 @@ async def _fetch_btc_usd_price() -> float:
     """Fetch the lowest BTC/USD price from multiple exchanges."""
     async with httpx.AsyncClient(timeout=30.0) as client:
         try:
-            tasks = [
-                asyncio.create_task(_kraken_btc_usd(client)),
-                asyncio.create_task(_coinbase_btc_usd(client)),
-                asyncio.create_task(_binance_btc_usdt(client)),
-            ]
-            valid_prices: list[float] = []
-
-            for future in asyncio.as_completed(tasks):
-                price = await future
-                if price is not None:
-                    valid_prices.append(price)
-
-                if len(valid_prices) >= 2:
-                    break
-
-            for task in tasks:
-                if not task.done():
-                    task.cancel()
-
+            prices = await asyncio.gather(
+                _kraken_btc_usd(client),
+                _coinbase_btc_usd(client),
+                _binance_btc_usdt(client),
+            )
+            valid_prices = [price for price in prices if price is not None]
             if not valid_prices:
                 logger.error("No valid BTC prices obtained from any exchange")
                 raise ValueError("Unable to fetch BTC price from any exchange")
-
             return min(valid_prices)
         except Exception as e:
             logger.error(

routstr/proxy.py

@@ -16,6 +16,7 @@ from .core.db import (
     create_session,
     get_session,
 )
+from .core.settings import settings
 from .payment.helpers import (
     calculate_discounted_max_cost,
     check_token_balance,
@@ -25,6 +26,7 @@ from .payment.helpers import (
 from .payment.models import Model
 from .upstream import BaseUpstreamProvider
 from .upstream.helpers import init_upstreams
+from .wallet import deserialize_token_from_string
 
 logger = get_logger(__name__)
 proxy_router = APIRouter()
@@ -96,8 +98,6 @@ async def refresh_model_maps() -> None:
     disabled_model_ids: set[str] = set()
 
     for provider in provider_rows:
-        if not provider.enabled:
-            continue
         for model in provider.models:
             if model.enabled:
                 overrides_by_id[model.id] = (model, provider.provider_fee)
@@ -148,6 +148,24 @@ async def proxy(
     else:
         model_id = request_body_dict.get("model", "unknown")
 
+    if "https://testnut.cashu.space" in settings.cashu_mints:
+        try:
+            token_str = None
+            if x_cashu_header := headers.get("x-cashu"):
+                token_str = x_cashu_header
+            elif auth_header := headers.get("authorization"):
+                parts = auth_header.split(" ")
+                if len(parts) > 1 and not parts[1].startswith("sk-"):
+                    token_str = parts[1]
+
+            if token_str:
+                token_obj = deserialize_token_from_string(token_str)
+                if token_obj.mint == "https://testnut.cashu.space":
+                    model_id = "mock/gpt-420-mock"
+                    request_body_dict["model"] = model_id
+        except Exception:
+            pass
+
     model_obj = get_model_instance(model_id)
     if not model_obj:
         return create_error_response(
@@ -202,42 +220,28 @@ async def proxy(
 
     headers = upstream.prepare_headers(dict(request.headers))
 
-    try:
-        if is_responses_api:
-            response = await upstream.forward_responses_request(
-                request,
-                path,
-                headers,
-                request_body,
-                key,
-                max_cost_for_model,
-                session,
-                model_obj,
-            )
-        else:
-            response = await upstream.forward_request(
-                request,
-                path,
-                headers,
-                request_body,
-                key,
-                max_cost_for_model,
-                session,
-                model_obj,
-            )
-    except Exception as e:
-        logger.error(
-            "Upstream request failed, ensuring payment is reverted",
-            extra={
-                "error": str(e),
-                "error_type": type(e).__name__,
-                "path": path,
-                "key_hash": key.hashed_key[:8] + "...",
-                "max_cost_for_model": max_cost_for_model,
-            },
+    if is_responses_api:
+        response = await upstream.forward_responses_request(
+            request,
+            path,
+            headers,
+            request_body,
+            key,
+            max_cost_for_model,
+            session,
+            model_obj,
+        )
+    else:
+        response = await upstream.forward_request(
+            request,
+            path,
+            headers,
+            request_body,
+            key,
+            max_cost_for_model,
+            session,
+            model_obj,
         )
-        await revert_pay_for_request(key, session, max_cost_for_model)
-        raise
 
     if response.status_code != 200:
         await revert_pay_for_request(key, session, max_cost_for_model)

routstr/upstream/base.py

@@ -12,7 +12,7 @@ from fastapi import BackgroundTasks, HTTPException, Request
 from fastapi.responses import Response, StreamingResponse
 from pydantic import BaseModel
 
-from ..auth import adjust_payment_for_tokens, revert_pay_for_request
+from ..auth import adjust_payment_for_tokens
 from ..core import get_logger
 from ..core.db import ApiKey, AsyncSession, create_session
 
@@ -1064,6 +1064,170 @@ class BaseUpstreamProvider:
                     },
                 )
 
+    async def handle_streaming_messages_completion(
+        self, response: httpx.Response, key: ApiKey, max_cost_for_model: int
+    ) -> StreamingResponse:
+        async def stream_with_cost(
+            max_cost_for_model: int,
+        ) -> AsyncGenerator[bytes, None]:
+            stored_chunks: list[bytes] = []
+            usage_finalized: bool = False
+            last_model_seen: str | None = None
+            input_tokens: int = 0
+            output_tokens: int = 0
+
+            async def finalize_without_usage() -> bytes | None:
+                nonlocal usage_finalized
+                if usage_finalized:
+                    return None
+                async with create_session() as new_session:
+                    fresh_key = await new_session.get(key.__class__, key.hashed_key)
+                    if not fresh_key:
+                        usage_finalized = True
+                        return None
+                    try:
+                        fallback: dict = {
+                            "model": last_model_seen or "unknown",
+                            "usage": None,
+                        }
+                        cost_data = await adjust_payment_for_tokens(
+                            fresh_key, fallback, new_session, max_cost_for_model
+                        )
+                        usage_finalized = True
+                        return f"event: cost\ndata: {json.dumps({'cost': cost_data})}\n\n".encode()
+                    except Exception:
+                        usage_finalized = True
+                        return None
+
+            try:
+                async for chunk in response.aiter_bytes():
+                    stored_chunks.append(chunk)
+                    try:
+                        decoded_chunk = chunk.decode("utf-8", errors="ignore")
+                        for line in decoded_chunk.split("\n"):
+                            if line.startswith("data: "):
+                                try:
+                                    data = json.loads(line[6:])
+                                    if isinstance(data, dict):
+                                        msg = data.get("message", {})
+                                        if msg and msg.get("model"):
+                                            last_model_seen = str(msg.get("model"))
+
+                                        if usage := msg.get("usage"):
+                                            input_tokens += usage.get("input_tokens", 0)
+                                            output_tokens += usage.get(
+                                                "output_tokens", 0
+                                            )
+
+                                        if usage := data.get("usage"):
+                                            input_tokens += usage.get("input_tokens", 0)
+                                            output_tokens += usage.get(
+                                                "output_tokens", 0
+                                            )
+                                except json.JSONDecodeError:
+                                    pass
+                    except Exception:
+                        pass
+
+                    yield chunk
+
+                usage_data = {
+                    "input_tokens": input_tokens,
+                    "output_tokens": output_tokens,
+                }
+
+                if input_tokens > 0 or output_tokens > 0:
+                    async with create_session() as new_session:
+                        fresh_key = await new_session.get(key.__class__, key.hashed_key)
+                        if fresh_key:
+                            try:
+                                combined_data = {
+                                    "model": last_model_seen or "unknown",
+                                    "usage": usage_data,
+                                }
+                                cost_data = await adjust_payment_for_tokens(
+                                    fresh_key,
+                                    combined_data,
+                                    new_session,
+                                    max_cost_for_model,
+                                )
+                                usage_finalized = True
+                                yield f"event: cost\ndata: {json.dumps({'cost': cost_data})}\n\n".encode()
+                            except Exception:
+                                pass
+
+                if not usage_finalized:
+                    maybe_cost_event = await finalize_without_usage()
+                    if maybe_cost_event is not None:
+                        yield maybe_cost_event
+
+            except Exception:
+                if not usage_finalized:
+                    await finalize_without_usage()
+                raise
+            finally:
+                if not usage_finalized:
+                    await finalize_without_usage()
+
+        response_headers = dict(response.headers)
+        response_headers.pop("content-encoding", None)
+        response_headers.pop("content-length", None)
+
+        return StreamingResponse(
+            stream_with_cost(max_cost_for_model),
+            status_code=response.status_code,
+            headers=response_headers,
+        )
+
+    async def handle_non_streaming_messages_completion(
+        self,
+        response: httpx.Response,
+        key: ApiKey,
+        session: AsyncSession,
+        deducted_max_cost: int,
+        path: str,
+    ) -> Response:
+        try:
+            content = await response.aread()
+            response_json = json.loads(content)
+
+            if path.endswith("count_tokens") and "usage" not in response_json:
+                input_tokens = response_json.get("input_tokens", 0)
+                response_json["usage"] = {"input_tokens": input_tokens}
+
+            cost_data = await adjust_payment_for_tokens(
+                key, response_json, session, deducted_max_cost
+            )
+            response_json["cost"] = cost_data
+
+            allowed_headers = {
+                "content-type",
+                "cache-control",
+                "date",
+                "vary",
+                "access-control-allow-origin",
+                "access-control-allow-methods",
+                "access-control-allow-headers",
+                "access-control-allow-credentials",
+                "access-control-expose-headers",
+                "access-control-max-age",
+            }
+
+            response_headers = {
+                k: v
+                for k, v in response.headers.items()
+                if k.lower() in allowed_headers
+            }
+
+            return Response(
+                content=json.dumps(response_json).encode(),
+                status_code=response.status_code,
+                headers=response_headers,
+                media_type="application/json",
+            )
+        except Exception:
+            raise
+
     async def forward_request(
         self,
         request: Request,
@@ -1157,7 +1321,54 @@ class BaseUpstreamProvider:
                     await client.aclose()
                 return mapped_error
 
-            if path.endswith("chat/completions") or path.endswith("embeddings"):
+            if (
+                path.endswith("chat/completions")
+                or path.endswith("embeddings")
+                or path.endswith("messages")
+                or path.endswith("messages/count_tokens")
+            ):
+                if path.endswith("messages"):
+                    client_wants_streaming = False
+                    if request_body:
+                        try:
+                            request_data = json.loads(request_body)
+                            client_wants_streaming = request_data.get("stream", False)
+                        except json.JSONDecodeError:
+                            pass
+
+                    content_type = response.headers.get("content-type", "")
+                    upstream_is_streaming = "text/event-stream" in content_type
+                    is_streaming = client_wants_streaming and upstream_is_streaming
+
+                    if is_streaming and response.status_code == 200:
+                        result = await self.handle_streaming_messages_completion(
+                            response, key, max_cost_for_model
+                        )
+                        background_tasks = BackgroundTasks()
+                        background_tasks.add_task(response.aclose)
+                        background_tasks.add_task(client.aclose)
+                        result.background = background_tasks
+                        return result
+
+                    if response.status_code == 200:
+                        try:
+                            return await self.handle_non_streaming_messages_completion(
+                                response, key, session, max_cost_for_model, path
+                            )
+                        finally:
+                            await response.aclose()
+                            await client.aclose()
+
+                if path.endswith("messages/count_tokens"):
+                    if response.status_code == 200:
+                        try:
+                            return await self.handle_non_streaming_messages_completion(
+                                response, key, session, max_cost_for_model, path
+                            )
+                        finally:
+                            await response.aclose()
+                            await client.aclose()
+
                 if path.endswith("chat/completions"):
                     client_wants_streaming = False
                     if request_body:
@@ -1256,8 +1467,6 @@ class BaseUpstreamProvider:
                 },
             )
 
-            await revert_pay_for_request(key, session, max_cost_for_model)
-
             if isinstance(exc, httpx.ConnectError):
                 error_message = "Unable to connect to upstream service"
             elif isinstance(exc, httpx.TimeoutException):
@@ -1289,8 +1498,6 @@ class BaseUpstreamProvider:
                 },
             )
 
-            await revert_pay_for_request(key, session, max_cost_for_model)
-
             return create_error_response(
                 "internal_error",
                 "An unexpected server error occurred",
@@ -1469,8 +1676,6 @@ class BaseUpstreamProvider:
                 },
             )
 
-            await revert_pay_for_request(key, session, max_cost_for_model)
-
             if isinstance(exc, httpx.ConnectError):
                 error_message = "Unable to connect to upstream service"
             elif isinstance(exc, httpx.TimeoutException):
@@ -1502,8 +1707,6 @@ class BaseUpstreamProvider:
                 },
             )
 
-            await revert_pay_for_request(key, session, max_cost_for_model)
-
             return create_error_response(
                 "internal_error",
                 "An unexpected server error occurred",

routstr/upstream/fake.py

@@ -0,0 +1,265 @@
+import asyncio
+import json
+import random
+from typing import AsyncIterator
+
+from fastapi import Request
+from fastapi.responses import Response, StreamingResponse
+
+from ..core.db import ApiKey, AsyncSession
+from ..payment.models import Architecture, Model, Pricing
+from .base import BaseUpstreamProvider
+
+
+class MockUpstreamProvider(BaseUpstreamProvider):
+    """Fack Mock Upstream provider specifically for Testing."""
+
+    provider_type = "mock"
+
+    async def forward_request(
+        self,
+        request: Request,
+        path: str,
+        headers: dict,
+        request_body: bytes | None,
+        key: ApiKey,
+        max_cost_for_model: int,
+        session: AsyncSession,
+        model_obj: Model,
+    ) -> Response | StreamingResponse:
+        if path.endswith("chat/completions"):
+            is_streaming = False
+            if request_body:
+                request_data = json.loads(request_body)
+                is_streaming = request_data.get("stream", False)
+
+            if is_streaming:
+
+                async def fake_streaming_response(
+                    chunk_size: int | None = None,
+                ) -> AsyncIterator[bytes]:
+                    suffix = random.randint(1000, 9999)
+                    req_id = f"gen-mock-stream-{suffix}"
+                    created = 1766138895
+                    model = "mock/gpt-420-mock"
+
+                    def make_chunk(
+                        delta: dict,
+                        finish_reason: str | None = None,
+                        usage: dict | None = None,
+                    ) -> bytes:
+                        chunk = {
+                            "id": req_id,
+                            "provider": "MockProvider",
+                            "model": model,
+                            "object": "chat.completion.chunk",
+                            "created": created,
+                            "choices": [
+                                {
+                                    "index": 0,
+                                    "delta": delta,
+                                    "finish_reason": finish_reason,
+                                    "native_finish_reason": "completed"
+                                    if finish_reason
+                                    else None,
+                                    "logprobs": None,
+                                }
+                            ],
+                        }
+                        if usage:
+                            chunk["usage"] = usage
+                        return f"data: {json.dumps(chunk)}\n\n".encode()
+
+                    # 1. Initial chunk
+                    yield make_chunk({"role": "assistant", "content": ""})
+                    await asyncio.sleep(0.02)
+
+                    # 2. Reasoning chunks
+                    reasoning_tokens = ["Mock", " reason", "ing", "..."]
+                    for token in reasoning_tokens:
+                        delta = {
+                            "role": "assistant",
+                            "content": "",
+                            "reasoning": token,
+                            "reasoning_details": [
+                                {
+                                    "type": "reasoning.summary",
+                                    "summary": token,
+                                    "format": "openai-responses-v1",
+                                    "index": 0,
+                                }
+                            ],
+                        }
+                        yield make_chunk(delta)
+                        await asyncio.sleep(0.03)
+
+                    # 3. Content chunks
+                    content_tokens = ["This", " is", " a", " mock", " stream", "."]
+                    for token in content_tokens:
+                        yield make_chunk({"role": "assistant", "content": token})
+                        await asyncio.sleep(0.03)
+
+                    # 4. Finish chunk
+                    yield make_chunk(
+                        {"role": "assistant", "content": ""}, finish_reason="stop"
+                    )
+
+                    # 5. Usage chunk
+                    usage_data = {
+                        "prompt_tokens": 10,
+                        "completion_tokens": 20,
+                        "total_tokens": 30,
+                        "cost": 0.001,
+                        "is_byok": False,
+                        "prompt_tokens_details": {
+                            "cached_tokens": 0,
+                            "audio_tokens": 0,
+                            "video_tokens": 0,
+                        },
+                        "cost_details": {
+                            "upstream_inference_cost": None,
+                            "upstream_inference_prompt_cost": 0,
+                            "upstream_inference_completions_cost": 0.001,
+                        },
+                        "completion_tokens_details": {
+                            "reasoning_tokens": 10,
+                            "image_tokens": 0,
+                        },
+                    }
+
+                    usage_chunk = {
+                        "id": req_id,
+                        "provider": "MockProvider",
+                        "model": model,
+                        "object": "chat.completion.chunk",
+                        "created": created,
+                        "choices": [
+                            {
+                                "index": 0,
+                                "delta": {"role": "assistant", "content": ""},
+                                "finish_reason": None,
+                                "native_finish_reason": None,
+                                "logprobs": None,
+                            }
+                        ],
+                        "usage": usage_data,
+                    }
+                    yield f"data: {json.dumps(usage_chunk)}\n\n".encode()
+
+                    # 6. DONE
+                    yield b"data: [DONE]\n\n"
+
+                    # 7. Cost
+                    cost_chunk = {
+                        "cost": {
+                            "base_msats": 0,
+                            "input_msats": 2,
+                            "output_msats": 10,
+                            "total_msats": 12,
+                        }
+                    }
+                    yield f"data: {json.dumps(cost_chunk)}\n\n".encode()
+
+                return StreamingResponse(
+                    fake_streaming_response(),
+                    200,
+                )
+
+            else:
+                suffix = random.randint(1000, 9999)
+                content_dict = {
+                    "id": f"gen-mock-{suffix}",
+                    "provider": "MockProvider",
+                    "model": "mock/gpt-5-mini",
+                    "object": "chat.completion",
+                    "created": 1766138655,
+                    "choices": [
+                        {
+                            "logprobs": None,
+                            "finish_reason": "length",
+                            "native_finish_reason": "max_output_tokens",
+                            "index": 0,
+                            "message": {
+                                "role": "assistant",
+                                "content": f"Mock Content {suffix}",
+                                "refusal": None,
+                                "reasoning": f"Mock Reasoning {suffix}",
+                                "reasoning_details": [
+                                    {
+                                        "format": "openai-responses-v1",
+                                        "index": 0,
+                                        "type": "reasoning.summary",
+                                        "summary": f"Mock Summary {suffix}",
+                                    },
+                                    {
+                                        "id": f"rs_mock_{suffix}",
+                                        "format": "openai-responses-v1",
+                                        "index": 0,
+                                        "type": "reasoning.encrypted",
+                                        "data": "mock_encrypted_data",
+                                    },
+                                ],
+                            },
+                        }
+                    ],
+                    "usage": {
+                        "prompt_tokens": 10,
+                        "completion_tokens": 10,
+                        "total_tokens": 20,
+                        "cost": 0,
+                        "is_byok": False,
+                        "prompt_tokens_details": {
+                            "cached_tokens": 0,
+                            "audio_tokens": 0,
+                            "video_tokens": 0,
+                        },
+                        "cost_details": {
+                            "upstream_inference_cost": None,
+                            "upstream_inference_prompt_cost": 0,
+                            "upstream_inference_completions_cost": 0,
+                        },
+                        "completion_tokens_details": {
+                            "reasoning_tokens": 5,
+                            "image_tokens": 0,
+                        },
+                    },
+                    "cost": {
+                        "base_msats": 0,
+                        "input_msats": 0,
+                        "output_msats": 0,
+                        "total_msats": 0,
+                    },
+                }
+                return Response(json.dumps(content_dict).encode(), 200)
+
+        elif path.endswith("embeddings"):
+            raise NotImplementedError
+        elif path.endswith("responses"):
+            raise NotImplementedError
+        else:
+            raise NotImplementedError
+
+    async def fetch_models(self) -> list[Model]:
+        return [
+            Model(
+                id="mock/gpt-420-mock",
+                name="mock/gpt-420-mock",
+                created=0,
+                description="mock model for testing",
+                context_length=8192,
+                architecture=Architecture(
+                    modality="text",
+                    input_modalities=["text"],
+                    output_modalities=["text"],
+                    tokenizer="",
+                    instruct_type=None,
+                ),
+                pricing=Pricing(prompt=0.01, completion=0.01),
+            ),
+        ]
+
+    def transform_model_name(self, model_id: str) -> str:
+        return "fake-model"
+
+    async def get_balance(self) -> float | None:
+        return 420.69

routstr/upstream/helpers.py

@@ -102,8 +102,6 @@ async def get_all_models_with_overrides(
             )
             for row in override_rows
             if row.upstream_provider_id is not None
-            and row.upstream_provider_id in providers_by_id
-            and providers_by_id[row.upstream_provider_id].enabled
         }
 
     all_models: dict[str, Model] = {}
@@ -220,6 +218,14 @@ async def init_upstreams() -> list[BaseUpstreamProvider]:
         results = await asyncio.gather(*tasks)
         upstreams = [p for p in results if p is not None]
 
+        if "https://testnut.cashu.space" in settings.cashu_mints:
+            from .fake import MockUpstreamProvider
+
+            mock_provider = MockUpstreamProvider("mock", "mock")
+            await mock_provider.refresh_models_cache()
+            upstreams.append(mock_provider)
+            logger.info("Initialized MockUpstreamProvider for testnut mint")
+
         return upstreams
 
 

routstr/wallet.py

@@ -313,6 +313,8 @@ async def periodic_payout() -> None:
         try:
             async with db.create_session() as session:
                 for mint_url in settings.cashu_mints:
+                    if mint_url == "https://testnut.cashu.space":
+                        continue
                     for unit in ["sat", "msat"]:
                         wallet = await get_wallet(mint_url, unit)
                         proofs = get_proofs_per_mint_and_unit(

tests/integration/test_example.py

@@ -10,6 +10,7 @@ from httpx import AsyncClient
 
 from .utils import (
     CashuTokenGenerator,
+    PerformanceValidator,
     ResponseValidator,
 )
 
@@ -158,7 +159,30 @@ async def test_error_handling(
     assert response.status_code == 401
 
 
+@pytest.mark.integration
+@pytest.mark.asyncio
+async def test_performance_requirements(integration_client: AsyncClient) -> None:
+    """Test that endpoints meet performance requirements"""
 
+    validator = PerformanceValidator()
+
+    # Test info endpoint performance
+    for i in range(50):
+        start = validator.start_timing("info_endpoint")
+        response = await integration_client.get("/")
+        validator.end_timing("info_endpoint", start)
+        assert response.status_code == 200
+
+    # Validate 95th percentile is under 500ms
+    result = validator.validate_response_time(
+        "info_endpoint", max_duration=0.5, percentile=0.95
+    )
+
+    assert result["valid"], (
+        f"Performance requirement failed: "
+        f"95th percentile was {result['percentile_time']:.3f}s "
+        f"(required < {result['max_allowed']}s)"
+    )
 
 
 @pytest.mark.integration

tests/integration/test_performance_load.py

@@ -9,7 +9,6 @@ import gc
 import statistics
 import time
 from typing import Any, Dict, List
-from unittest.mock import patch
 
 import psutil
 import pytest
@@ -106,46 +105,42 @@ class TestPerformanceBaseline:
             ("GET", "/v1/wallet/info", authenticated_client, None),
         ]
 
-        # Enable provider discovery for this test
-        with patch(
-            "routstr.core.settings.settings.providers_refresh_interval_seconds", 300
-        ):
-            # Warm up
-            for _ in range(10):
-                await integration_client.get("/")
+        # Warm up
+        for _ in range(10):
+            await integration_client.get("/")
 
-            # Test each endpoint
-            for method, path, client, data in endpoints:
-                response_times = []
+        # Test each endpoint
+        for method, path, client, data in endpoints:
+            response_times = []
 
-                for i in range(100):
-                    start = time.time()
+            for i in range(100):
+                start = time.time()
 
-                    if method == "GET":
-                        response = await client.get(path)
-                    else:
-                        response = await client.post(path, json=data)
+                if method == "GET":
+                    response = await client.get(path)
+                else:
+                    response = await client.post(path, json=data)
 
-                    duration = time.time() - start
-                    response_times.append(duration * 1000)  # Convert to ms
+                duration = time.time() - start
+                response_times.append(duration * 1000)  # Convert to ms
 
-                    assert response.status_code in [200, 201]
+                assert response.status_code in [200, 201]
 
-                    if i % 10 == 0:
-                        metrics.record_system_metrics()
+                if i % 10 == 0:
+                    metrics.record_system_metrics()
 
-                # Verify 95th percentile < 500ms
-                p95 = sorted(response_times)[int(len(response_times) * 0.95)]
-                assert p95 < 500, (
-                    f"{method} {path} p95 response time {p95}ms exceeds 500ms limit"
-                )
+            # Verify 95th percentile < 500ms
+            p95 = sorted(response_times)[int(len(response_times) * 0.95)]
+            assert p95 < 500, (
+                f"{method} {path} p95 response time {p95}ms exceeds 500ms limit"
+            )
 
-                print(f"\n{method} {path}:")
-                print(f"  Mean: {statistics.mean(response_times):.2f}ms")
-                print(f"  P95: {p95:.2f}ms")
-                print(
-                    f"  P99: {sorted(response_times)[int(len(response_times) * 0.99)]:.2f}ms"
-                )
+            print(f"\n{method} {path}:")
+            print(f"  Mean: {statistics.mean(response_times):.2f}ms")
+            print(f"  P95: {p95:.2f}ms")
+            print(
+                f"  P99: {sorted(response_times)[int(len(response_times) * 0.99)]:.2f}ms"
+            )
 
 
 @pytest.mark.integration

tests/integration/test_provider_management.py

@@ -3,7 +3,7 @@ Integration tests for provider management functionality.
 Tests GET /v1/providers/ endpoint for listing and managing providers.
 """
 
-from typing import Any, Generator
+from typing import Any
 from unittest.mock import patch
 
 import pytest
@@ -11,7 +11,7 @@ from httpx import AsyncClient
 
 from routstr.discovery import _PROVIDERS_CACHE
 
-from .utils import ResponseValidator
+from .utils import PerformanceValidator, ResponseValidator
 
 
 @pytest.fixture(autouse=True)
@@ -19,15 +19,6 @@ def _clear_providers_cache() -> None:
     _PROVIDERS_CACHE.clear()
 
 
-@pytest.fixture(autouse=True)
-def _enable_provider_discovery() -> Generator[None, Any, Any]:
-    """Enable provider discovery for all tests in this module"""
-    with patch(
-        "routstr.core.settings.settings.providers_refresh_interval_seconds", 300
-    ):
-        yield
-
-
 @pytest.mark.integration
 @pytest.mark.asyncio
 async def test_providers_endpoint_default_response(
@@ -527,6 +518,46 @@ async def test_providers_endpoint_response_format(
             assert isinstance(data_json["providers"], list)
 
 
+@pytest.mark.integration
+@pytest.mark.asyncio
+async def test_providers_endpoint_performance(integration_client: AsyncClient) -> None:
+    """Test providers endpoint meets performance requirements"""
+
+    # Mock quick responses to avoid network delays
+    mock_events: list[dict[str, Any]] = [
+        {
+            "id": f"event{i}",
+            "content": f"Provider: http://provider{i}.onion",
+            "created_at": 1234567890 + i,
+        }
+        for i in range(5)
+    ]
+
+    validator = PerformanceValidator()
+
+    with patch(
+        "routstr.discovery.query_nostr_relay_for_providers", return_value=mock_events
+    ):
+        with patch("routstr.discovery.fetch_provider_health") as mock_fetch:
+            mock_fetch.return_value = {"status_code": 200, "json": {"status": "online"}}
+
+            # Test multiple requests
+            for i in range(10):
+                start = validator.start_timing("providers_endpoint")
+                response = await integration_client.get("/v1/providers/")
+                validator.end_timing("providers_endpoint", start)
+
+                assert response.status_code == 200
+
+    # Validate performance (should be fast with mocked dependencies)
+    perf_result = validator.validate_response_time(
+        "providers_endpoint",
+        max_duration=2.0,  # Allow more time since it involves multiple operations
+        percentile=0.95,
+    )
+    assert perf_result["valid"], f"Performance requirement failed: {perf_result}"
+
+
 @pytest.mark.integration
 @pytest.mark.asyncio
 async def test_providers_endpoint_concurrent_requests(

tests/integration/test_proxy_get_endpoints.py

@@ -18,6 +18,7 @@ from routstr.core.db import ApiKey
 
 from .utils import (
     ConcurrencyTester,
+    PerformanceValidator,
 )
 
 
@@ -550,7 +551,39 @@ async def test_proxy_get_concurrent_requests(
             assert response.status_code == 200
 
 
+@pytest.mark.integration
+@pytest.mark.asyncio
+async def test_proxy_get_performance_requirements(
+    integration_client: AsyncClient, authenticated_client: AsyncClient
+) -> None:
+    """Test that GET proxy requests meet performance requirements"""
 
+    validator = PerformanceValidator()
+
+    with patch("httpx.AsyncClient.request") as mock_request:
+        mock_response = AsyncMock()
+        mock_response.status_code = 200
+        mock_response.headers = {"content-type": "application/json"}
+        mock_response.json = MagicMock(return_value={"performance": "test"})
+        mock_response.text = '{"performance": "test"}'
+        mock_response.iter_bytes = AsyncMock(return_value=[b'{"performance": "test"}'])
+        mock_request.return_value = mock_response
+
+        # Test multiple requests for performance measurement
+        for i in range(20):
+            start = validator.start_timing("proxy_get")
+            response = await authenticated_client.get(f"/v1/perf-test-{i}")
+            validator.end_timing("proxy_get", start)
+
+            assert response.status_code == 200
+
+    # Validate performance requirements
+    perf_result = validator.validate_response_time(
+        "proxy_get",
+        max_duration=1.0,  # Should complete within 1 second
+        percentile=0.95,
+    )
+    assert perf_result["valid"], f"Performance requirement failed: {perf_result}"
 
 
 @pytest.mark.integration

tests/integration/test_proxy_post_endpoints.py

@@ -15,6 +15,7 @@ from httpx import ASGITransport, AsyncClient
 
 from .utils import (
     ConcurrencyTester,
+    PerformanceValidator,
 )
 
 
@@ -289,7 +290,55 @@ async def test_proxy_post_unauthorized_access(integration_client: AsyncClient) -
     assert response.status_code in [400, 401]
 
 
+@pytest.mark.integration
+@pytest.mark.asyncio
+async def test_proxy_post_performance(
+    integration_client: AsyncClient, authenticated_client: AsyncClient
+) -> None:
+    """Test POST endpoint performance requirements"""
 
+    test_payload = {
+        "model": "gpt-3.5-turbo",
+        "messages": [{"role": "user", "content": "Performance test"}],
+    }
+
+    validator = PerformanceValidator()
+
+    with patch("httpx.AsyncClient.send") as mock_send:
+        # Mock fast responses
+        async def mock_iter_bytes(*args: Any, **kwargs: Any) -> Any:
+            yield b'{"choices": [{"message": {"content": "Fast"}}], "usage": {"total_tokens": 5}}'
+
+        mock_response = AsyncMock()
+        mock_response.status_code = 200
+        mock_response.headers = {"content-type": "application/json"}
+        response_data = {
+            "choices": [{"message": {"content": "Fast"}}],
+            "usage": {"total_tokens": 5},
+        }
+        mock_response.text = json.dumps(response_data)
+        mock_response.json = AsyncMock(return_value=response_data)
+        mock_response.iter_bytes = mock_iter_bytes
+        mock_response.aiter_bytes = mock_iter_bytes
+        mock_send.return_value = mock_response
+
+        # Run multiple requests for performance measurement
+        for i in range(20):
+            start = validator.start_timing("proxy_post")
+            response = await authenticated_client.post(
+                "/v1/chat/completions", json=test_payload
+            )
+            validator.end_timing("proxy_post", start)
+
+            assert response.status_code == 200
+
+    # Validate performance
+    perf_result = validator.validate_response_time(
+        "proxy_post",
+        max_duration=1.5,  # Allow slightly more time for POST
+        percentile=0.95,
+    )
+    assert perf_result["valid"], f"Performance requirement failed: {perf_result}"
 
 
 @pytest.mark.integration

tests/integration/test_wallet_information.py

@@ -3,7 +3,7 @@ Integration tests for wallet information retrieval endpoints.
 Tests GET /v1/wallet/ and GET /v1/wallet/info endpoints with various scenarios.
 """
 
-
+import time
 from datetime import datetime, timedelta
 from typing import Any
 
@@ -367,4 +367,30 @@ async def test_wallet_info_with_special_characters_in_headers(
     # Note: Current implementation doesn't return refund_address in response
 
 
+@pytest.mark.integration
+@pytest.mark.asyncio
+@pytest.mark.slow
+async def test_wallet_endpoints_performance(authenticated_client: AsyncClient) -> None:
+    """Test wallet endpoints meet performance requirements"""
 
+    # Warm up
+    await authenticated_client.get("/v1/wallet/")
+
+    # Measure response times
+    response_times = []
+
+    for _ in range(50):
+        start_time = time.time()
+        response = await authenticated_client.get("/v1/wallet/")
+        end_time = time.time()
+
+        assert response.status_code == 200
+        response_times.append(end_time - start_time)
+
+    # Calculate statistics
+    avg_time = sum(response_times) / len(response_times)
+    max_time = max(response_times)
+
+    # Performance assertions
+    assert avg_time < 0.1  # Average should be under 100ms
+    assert max_time < 0.5  # No request should take more than 500ms

tests/integration/test_wallet_refund.py

@@ -537,4 +537,42 @@ async def test_refund_with_expired_key(
         assert response.json()["recipient"] == "expired@ln.address"
 
 
+@pytest.mark.integration
+@pytest.mark.asyncio
+@pytest.mark.slow
+async def test_refund_performance(
+    integration_client: AsyncClient, testmint_wallet: Any
+) -> None:
+    """Test refund endpoint performance"""
 
+    import time
+
+    # Create multiple API keys
+    api_keys = []
+    for i in range(10):
+        token = await testmint_wallet.mint_tokens(100 + i)
+        # Use cashu token as Bearer auth to create API key
+        integration_client.headers["Authorization"] = f"Bearer {token}"
+        response = await integration_client.get("/v1/wallet/info")
+        assert response.status_code == 200
+        api_keys.append(response.json()["api_key"])
+
+    # Measure refund times
+    refund_times = []
+
+    for api_key in api_keys:
+        integration_client.headers["Authorization"] = f"Bearer {api_key}"
+
+        start_time = time.time()
+        response = await integration_client.post("/v1/wallet/refund")
+        end_time = time.time()
+
+        assert response.status_code == 200
+        refund_times.append(end_time - start_time)
+
+    # Performance assertions
+    avg_time = sum(refund_times) / len(refund_times)
+    max_time = max(refund_times)
+
+    assert avg_time < 0.5  # Average under 500ms
+    assert max_time < 1.0  # No refund takes more than 1 second