fishr

FREE LLM ACCESS FROM PYTHON. NO KEYS! NO AUTH! NO LIMITS!

Install

pip install fishr

Quick Start

from fishr import Client

client = Client()

# Chat completion
response = client.chat.completions.create(
    model="noxus/openai",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.text)

# Streaming
response = client.chat.completions.create(
    model="noxus/openai",
    messages=[{"role": "user", "content": "Tell me a story"}],
    stream=True,
)
for chunk in response:
    print(chunk.choices[0].delta.content or "", end="")

# Web search
response = client.chat.completions.create(
    model="noxus/grok-4.3",
    messages=[{"role": "user", "content": "Latest news on SpaceX"}],
    web_search=True,
)
print(response.text)

# Image generation
result = client.images.generate(
    model="deepai/image",
    prompt="A cat riding a skateboard",
)
print(result.data[0].url)

# Text-to-speech (make/* models — 39 voices)
result = client.audio.speech.create(
    model="make/aura",
    input="Hello from fishr.",
)
with open("out.wav", "wb") as f:
    f.write(result.data[0].audio)

# Agent
result = client.agents.run(
    model="noxus/openai",
    prompt="Research quantum computing",
    tools=[{"type": "web_search"}],
)
print(result.content)

# Tool calling (kai/* models support native OpenAI-style tools)
tools = [{
    "type": "function",
    "function": {
        "name": "get_weather",
        "description": "Get weather for a location",
        "parameters": {
            "type": "object",
            "properties": {"location": {"type": "string"}},
            "required": ["location"],
        },
    },
}]
resp = client.chat.completions.create(
    model="kai/auto",
    messages=[{"role": "user", "content": "What is the weather in Tokyo?"}],
    tools=tools,
)
for tc in resp.choices[0].message.tool_calls:
    print(tc.function.name, tc.function.arguments)

# Multi-turn conversation
conv = client.conversation(model="noxus/openai")
conv.system("You are a helpful assistant.")
conv.ask("Remember: 42")
conv.ask("What number did I say?")

Async

from fishr import AsyncClient

client = AsyncClient()

response = await client.chat.completions.create(
    model="deepai/standard",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.text)

# Async streaming
response = await client.chat.completions.create(
    model="opera/aria",
    messages=[{"role": "user", "content": "Tell me a story"}],
    stream=True,
)
async for chunk in response:
    print(chunk.choices[0].delta.content or "", end="")

# Async conversation
conv = client.conversation(model="noxus/openai")
await conv.ask("Remember: 42")
await conv.ask("What number did I say?")

Direct Provider Access

from fishr import OperaAria, Yqcloud

# Use a provider directly
opera = OperaAria()
result = opera.ask("Hello!", model="opera/aria")
print(result.content)

yqcloud = Yqcloud()
result = yqcloud.chat([{"role": "user", "content": "Hi!"}])
print(result.content)

Models

Models are specified as provider/name:

Provider	Models	Features
Noxus	noxus/openai, noxus/google, noxus/sonnet-4.6, noxus/sonnet-3.5, noxus/grok-4.3, noxus/perplexity, noxus/metaai, noxus/qwen	Web search, images, history, system messages
DeepAI	deepai/standard, deepai/online, deepai/gemma-4, deepai/gemini-2.5-flash-lite, deepai/deepseek-v3.2, deepai/image	Web search (online), images, history, system messages
Quillbot	quillbot/quillbot, quillbot/quillbot-search	Web search (quillbot-search), history, system messages
NoTrack	notrack/fast, notrack/standard, notrack/reasoning	History
DphnAI	dphnai/24b, dphnai/6b	History, system messages
Yqcloud	yqcloud/gpt-4	History, system messages
Opera Aria	opera/aria	Images, history, system messages
Kai	kai/auto, kai/m.1, kai/xs-2.1, kai/xs.2, kai/north-mini, kai/nemo3-ultra, kai/nemo3-super, kai/nemo3-nemo, kai/3.7-flash, kai/openfree	Tool calling, history, system messages (some support images)
Make	make/aura, make/breeze, make/cypress, make/drift, make/echo, make/flare, make/gem, make/hazel, make/ivy, make/jazz, make/kite, make/lumen, make/mist, make/saffron, make/solstice, make/pearl, make/quartz, make/ripple, make/cobalt, make/tide, make/vale, make/wren, make/ash, make/brook, make/cedar, make/dawn, make/fern, make/glen, make/harbor, make/indigo, make/juniper, make/lotus, make/maple, make/nettle, make/opal, make/pine, make/river, make/slate, make/willow	TTS (39 voices, WAV)

The provider prefix is optional for Noxus (default).

Requirements

• Python >= 3.11