groq-chain 1.0.0

Dead-simple Groq LLM chaining. No bloat.

groq-chain

Dead-simple Groq LLM chaining. One dependency. No bloat.

LangChain is overkill for most things. groq-chain gives you prompt chaining in plain Python — pass the output of one LLM call into the next, with zero magic.

---

Install

pip install groq-chain

Or from source:

git clone https://github.com/iamadhitya1/groq-chain
pip install -e groq-chain/

---

Quick start

from groqchain import GroqChain

chain = GroqChain(api_key="gsk_...")  # or set GROQ_API_KEY env var

# Single call
result = chain.run("Summarize this in 3 bullet points: {text}", text="...")
print(result)

---

Chained calls

Pass the output of each step into the next automatically:

result = (
    GroqChain(api_key="gsk_...")
    .step("Extract 3 key insights from: {text}", output_key="insights", text="...")
    .step("Write a LinkedIn post based on these insights: {insights}")
    .run()
)
print(result)

Each .step() receives the previous step's output via output_key.

---

Get all step outputs

results = (
    GroqChain(api_key="gsk_...")
    .step("Translate to French: {text}", output_key="french", text="Hello world")
    .step("Now translate the French to Spanish: {french}", output_key="spanish")
    .run_all()
)

print(results["french"])   # Bonjour le monde
print(results["spanish"])  # Hola mundo

---

Inject context

chain = (
    GroqChain(api_key="gsk_...")
    .context(language="Hindi", tone="casual")
    .step("Write a {tone} greeting in {language}")
)
result = chain.run()

---

System prompt

chain = GroqChain(
    api_key="gsk_...",
    system="You are a senior software engineer. Be concise and technical.",
)
result = chain.run("Review this code: {code}", code="...")

---

All options

GroqChain(
    api_key="gsk_...",                      # or GROQ_API_KEY env var
    model="llama-3.3-70b-versatile",        # any Groq model
    temperature=0.7,
    max_tokens=1024,
    system="Optional system prompt",
)

Available Groq models:

• llama-3.3-70b-versatile ← default
• llama-3.1-8b-instant
• mixtral-8x7b-32768
• gemma2-9b-it

---

Real-world example — document pipeline

import os
from groqchain import GroqChain

chain = GroqChain(api_key=os.environ["GROQ_API_KEY"])

with open("contract.txt") as f:
    doc = f.read()

results = (
    chain
    .step("Summarize this legal document: {doc}", output_key="summary", doc=doc)
    .step("List any risky clauses from this summary: {summary}", output_key="risks")
    .step("Rate the overall risk from 1-10 and explain why: {risks}", output_key="rating")
    .run_all()
)

print("Summary:", results["summary"])
print("Risks:",   results["risks"])
print("Rating:",  results["rating"])

---

License

MIT © 2025 M Adhitya

Built at Rewrite Labs — powering AI tools used by thousands.