Skip to content

Drop in a screenshot and convert it to clean code (HTML/Tailwind/React/Vue)

Python

seperate files

Abi Raja04b9b74e0468
diff --git a/backend/agent/providers/anthropic.py b/backend/agent/providers/anthropic.py
index 0e10769b..05841042 100644
--- a/backend/agent/providers/anthropic.py
+++ b/backend/agent/providers/anthropic.py
@@ -13,14 +13,14 @@ from openai.types.chat import ChatCompletionMessageParam
from PIL import Image
from agent.providers.base import (
- MODEL_PRICING,
EventSink,
ExecutedToolCall,
ProviderSession,
ProviderTurn,
StreamEvent,
- TokenUsage,
)
+from agent.providers.pricing import MODEL_PRICING
+from agent.providers.token_usage import TokenUsage
from agent.tools import CanonicalToolDefinition, ToolCall, parse_json_arguments
from llm import Llm
diff --git a/backend/agent/providers/base.py b/backend/agent/providers/base.py
index 2274522e..604cc50c 100644
--- a/backend/agent/providers/base.py
+++ b/backend/agent/providers/base.py
@@ -1,5 +1,5 @@
from dataclasses import dataclass
-from typing import Any, Awaitable, Callable, Dict, Literal, Optional, Protocol
+from typing import Any, Awaitable, Callable, Literal, Optional, Protocol
from agent.tools import ToolCall, ToolExecutionResult
@@ -20,103 +20,6 @@ class StreamEvent:
tool_arguments: Any = None
-@dataclass
-class TokenUsage:
- """Unified token usage across all providers.
-
- Log line example:
- [TOKEN USAGE] provider=gemini model=... | input=1000 output=500
- cache_read=200 cache_write=0 total=1700 cost=$0.0020
-
- Fields:
- input: Non-cached input tokens (billed at full input rate).
- For providers whose API includes cached tokens in the
- prompt count (OpenAI, Gemini), cached tokens are
- subtracted so this is always *exclusive* of cache_read.
- output: Output tokens including thinking/reasoning (billed at
- output rate).
- cache_read: Input tokens served from cache (billed at reduced rate).
- cache_write: Input tokens written to cache (Anthropic only).
- total: All tokens as reported by the provider API. Equals
- input + cache_read + output (+ thinking for Gemini).
-
- Total input sent to the model = input + cache_read.
- Cost = (input * input_rate + output * output_rate
- + cache_read * cache_read_rate + cache_write * cache_write_rate)
- / 1_000_000
- """
-
- input: int = 0
- output: int = 0
- cache_read: int = 0
- cache_write: int = 0
- total: int = 0
-
- def accumulate(self, other: "TokenUsage") -> None:
- self.input += other.input
- self.output += other.output
- self.cache_read += other.cache_read
- self.cache_write += other.cache_write
- self.total += other.total
-
- def cost(self, pricing: "ModelPricing") -> float:
- """Compute cost in USD using per-million-token rates."""
- return (
- self.input * pricing.input
- + self.output * pricing.output
- + self.cache_read * pricing.cache_read
- + self.cache_write * pricing.cache_write
- ) / 1_000_000
-
-
-@dataclass
-class ModelPricing:
- """Per-million-token pricing in USD."""
-
- input: float = 0.0
- output: float = 0.0
- cache_read: float = 0.0
- cache_write: float = 0.0
-
-
-# Pricing keyed by the API model name string sent to the provider.
-MODEL_PRICING: Dict[str, ModelPricing] = {
- # --- OpenAI ---
- "gpt-4.1-2025-04-14": ModelPricing(
- input=2.00, output=8.00, cache_read=0.50
- ),
- "gpt-5.2-codex": ModelPricing(
- input=1.75, output=14.00, cache_read=0.4375
- ),
- "gpt-5.3-codex": ModelPricing(
- input=1.75, output=14.00, cache_read=0.4375
- ),
- # --- Anthropic ---
- "claude-sonnet-4-6": ModelPricing(
- input=3.00, output=15.00, cache_read=0.30, cache_write=3.75
- ),
- "claude-sonnet-4-5-20250929": ModelPricing(
- input=3.00, output=15.00, cache_read=0.30, cache_write=3.75
- ),
- "claude-opus-4-5-20251101": ModelPricing(
- input=5.00, output=25.00, cache_read=0.50, cache_write=6.25
- ),
- "claude-opus-4-6": ModelPricing(
- input=5.00, output=25.00, cache_read=0.50, cache_write=6.25
- ),
- # --- Gemini ---
- "gemini-3-flash-preview": ModelPricing(
- input=0.50, output=3.00, cache_read=0.05
- ),
- "gemini-3-pro-preview": ModelPricing(
- input=2.00, output=12.00, cache_read=0.20
- ),
- "gemini-3.1-pro-preview": ModelPricing(
- input=2.00, output=12.00, cache_read=0.20
- ),
-}
-
-
@dataclass
class ProviderTurn:
assistant_text: str
diff --git a/backend/agent/providers/gemini.py b/backend/agent/providers/gemini.py
index 1c0c3b49..c803d201 100644
--- a/backend/agent/providers/gemini.py
+++ b/backend/agent/providers/gemini.py
@@ -10,14 +10,14 @@ from google.genai import types
from openai.types.chat import ChatCompletionMessageParam
from agent.providers.base import (
- MODEL_PRICING,
EventSink,
ExecutedToolCall,
ProviderSession,
ProviderTurn,
StreamEvent,
- TokenUsage,
)
+from agent.providers.pricing import MODEL_PRICING
+from agent.providers.token_usage import TokenUsage
from agent.tools import CanonicalToolDefinition, ToolCall
from llm import Llm
diff --git a/backend/agent/providers/openai.py b/backend/agent/providers/openai.py
index 91b56057..e1a3cdae 100644
--- a/backend/agent/providers/openai.py
+++ b/backend/agent/providers/openai.py
@@ -9,14 +9,14 @@ from openai import AsyncOpenAI
from openai.types.chat import ChatCompletionMessageParam
from agent.providers.base import (
- MODEL_PRICING,
EventSink,
ExecutedToolCall,
ProviderSession,
ProviderTurn,
StreamEvent,
- TokenUsage,
)
+from agent.providers.pricing import MODEL_PRICING
+from agent.providers.token_usage import TokenUsage
from agent.state import ensure_str
from agent.tools import CanonicalToolDefinition, ToolCall, parse_json_arguments
from llm import Llm, get_openai_api_name, get_openai_reasoning_effort
diff --git a/backend/agent/providers/pricing.py b/backend/agent/providers/pricing.py
new file mode 100644
index 00000000..6c469795
--- /dev/null
+++ b/backend/agent/providers/pricing.py
@@ -0,0 +1,40 @@
+from typing import Dict
+
+from agent.providers.token_usage import ModelPricing
+
+# Pricing keyed by the API model name string sent to the provider.
+MODEL_PRICING: Dict[str, ModelPricing] = {
+ # --- OpenAI ---
+ "gpt-4.1-2025-04-14": ModelPricing(
+ input=2.00, output=8.00, cache_read=0.50
+ ),
+ "gpt-5.2-codex": ModelPricing(
+ input=1.75, output=14.00, cache_read=0.4375
+ ),
+ "gpt-5.3-codex": ModelPricing(
+ input=1.75, output=14.00, cache_read=0.4375
+ ),
+ # --- Anthropic ---
+ "claude-sonnet-4-6": ModelPricing(
+ input=3.00, output=15.00, cache_read=0.30, cache_write=3.75
+ ),
+ "claude-sonnet-4-5-20250929": ModelPricing(
+ input=3.00, output=15.00, cache_read=0.30, cache_write=3.75
+ ),
+ "claude-opus-4-5-20251101": ModelPricing(
+ input=5.00, output=25.00, cache_read=0.50, cache_write=6.25
+ ),
+ "claude-opus-4-6": ModelPricing(
+ input=5.00, output=25.00, cache_read=0.50, cache_write=6.25
+ ),
+ # --- Gemini ---
+ "gemini-3-flash-preview": ModelPricing(
+ input=0.50, output=3.00, cache_read=0.05
+ ),
+ "gemini-3-pro-preview": ModelPricing(
+ input=2.00, output=12.00, cache_read=0.20
+ ),
+ "gemini-3.1-pro-preview": ModelPricing(
+ input=2.00, output=12.00, cache_read=0.20
+ ),
+}
diff --git a/backend/agent/providers/token_usage.py b/backend/agent/providers/token_usage.py
new file mode 100644
index 00000000..28ca55d8
--- /dev/null
+++ b/backend/agent/providers/token_usage.py
@@ -0,0 +1,62 @@
+from __future__ import annotations
+
+from dataclasses import dataclass
+
+
+@dataclass
+class ModelPricing:
+ """Per-million-token pricing in USD."""
+
+ input: float = 0.0
+ output: float = 0.0
+ cache_read: float = 0.0
+ cache_write: float = 0.0
+
+
+@dataclass
+class TokenUsage:
+ """Unified token usage across all providers.
+
+ Log line example:
+ [TOKEN USAGE] provider=gemini model=... | input=1000 output=500
+ cache_read=200 cache_write=0 total=1700 cost=$0.0020
+
+ Fields:
+ input: Non-cached input tokens (billed at full input rate).
+ For providers whose API includes cached tokens in the
+ prompt count (OpenAI, Gemini), cached tokens are
+ subtracted so this is always *exclusive* of cache_read.
+ output: Output tokens including thinking/reasoning (billed at
+ output rate).
+ cache_read: Input tokens served from cache (billed at reduced rate).
+ cache_write: Input tokens written to cache (Anthropic only).
+ total: All tokens as reported by the provider API. Equals
+ input + cache_read + output (+ thinking for Gemini).
+
+ Total input sent to the model = input + cache_read.
+ Cost = (input * input_rate + output * output_rate
+ + cache_read * cache_read_rate + cache_write * cache_write_rate)
+ / 1_000_000
+ """
+
+ input: int = 0
+ output: int = 0
+ cache_read: int = 0
+ cache_write: int = 0
+ total: int = 0
+
+ def accumulate(self, other: TokenUsage) -> None:
+ self.input += other.input
+ self.output += other.output
+ self.cache_read += other.cache_read
+ self.cache_write += other.cache_write
+ self.total += other.total
+
+ def cost(self, pricing: ModelPricing) -> float:
+ """Compute cost in USD using per-million-token rates."""
+ return (
+ self.input * pricing.input
+ + self.output * pricing.output
+ + self.cache_read * pricing.cache_read
+ + self.cache_write * pricing.cache_write
+ ) / 1_000_000
diff --git a/backend/tests/test_token_usage.py b/backend/tests/test_token_usage.py
index b896361f..c192d88c 100644
--- a/backend/tests/test_token_usage.py
+++ b/backend/tests/test_token_usage.py
@@ -2,7 +2,8 @@
from types import SimpleNamespace
-from agent.providers.base import MODEL_PRICING, ModelPricing, TokenUsage
+from agent.providers.pricing import MODEL_PRICING
+from agent.providers.token_usage import ModelPricing, TokenUsage
from agent.providers.anthropic import _extract_anthropic_usage
from agent.providers.gemini import _extract_usage as _extract_gemini_usage
from agent.providers.openai import _extract_openai_usage