Documentation

Colnma Introduction

Prompt Management • Prompt Experiments (LLM-as-Judge) • Context Packs (universal RAG pipelines). Powering AI with a brain — enabling true context-aware intelligence through prompts that think, remember, and reason.

What is Colnma?

Colnma is a unified platform for Prompt Management, Experiments, and Context Pipelines — powering AI with a brain for context‑aware responses. It helps developers and teams transform static prompts into intelligent, reusable, and governed systems that learn and evolve with your data.

Benefits

Quickstart Installation

npm install @Colnma/client
# or
pip install Colnma

JavaScript SDK

import { ColnmaClient } from "@Colnma/client";
const client = new ColnmaClient({ projectKey: "pv_sk_abc123" });
const systemPrompt = await client.getPrompt("hr-assistant", { question: "What is our PTO policy?" });
console.log(systemPrompt);

Python SDK

from Colnma import ColnmaClient
client = ColnmaClient(project_key="pv_sk_abc123")
system_prompt = client.get_prompt("hr-assistant", {"question": "What is our PTO policy?"})
print(system_prompt)

REST API

Request

POST /api/sdk/v1/prompt/client/{project_key}/{prompt_name}
Authorization: Bearer YOUR_TOKEN
Content-Type: application/json

{
  "variables": { "question": "What is our PTO policy?" }
}

Response

{
  "prompt": "You are \"Colnma AI,\" a world-class expert in prompt engineering and Large Language Model interaction. You firmly believe that a well-crafted prompt is the key to unlocking an AI's full potential.\nThe user has submitted a prompt for analysis. The content of their original prompt is provided in the eqeqed...\n##Use below Web References\n- https://colnma.com"
}

Responses: 200 OK → returns compiled system prompt; 401 Unauthorized; 422 Validation Error; 429 Rate limit exceeded.

Guides Overview

Colnma integrates with major agent-building frameworks. Below are examples for LangChain, LangChain.js, Vercel AI SDK, LlamaIndex, and HTTP-only flows.

LangChain (Python)

from Colnma import ColnmaClient
from langchain_openai import ChatOpenAI
from langchain.prompts import ChatPromptTemplate

ptv = ColnmaClient(project_key="pv_sk_abc123")
prompt_text = ptv.get_prompt("hr-assistant", {"question": "What is our PTO policy?"})

llm = ChatOpenAI(model="gpt-4o-mini")
prompt = ChatPromptTemplate.from_messages([
    ("system", prompt_text),
    ("user", "What is our PTO policy?")
])
result = (prompt | llm).invoke({})
print(result.content)

LangChain.js (TypeScript)

import { ColnmaClient } from "@Colnma/client";
import { ChatOpenAI } from "@langchain/openai";
import { ChatPromptTemplate } from "@langchain/core/prompts";

const ptv = new ColnmaClient({ projectKey: "pv_sk_abc123" });
const promptText = await ptv.getPrompt("hr-assistant", { question: "What is our PTO policy?" });

const model = new ChatOpenAI({ modelName: "gpt-4o-mini" });
const prompt = ChatPromptTemplate.fromMessages([
  ["system", promptText],
  ["user", "What is our PTO policy?"],
]);
const res = await prompt.pipe(model).invoke({});
console.log(res.content);

Vercel AI SDK (JavaScript)

import { ColnmaClient } from "@Colnma/client";
import { generateText } from "ai";
import { openai } from "@ai-sdk/openai";

const ptv = new ColnmaClient({ projectKey: "pv_sk_abc123" });
const systemPrompt = await ptv.getPrompt("hr-assistant", { question: "What is our PTO policy?" });

const { text } = await generateText({
  model: openai("gpt-4o-mini"),
  system: systemPrompt,
  prompt: "What is our PTO policy?",
});
console.log(text);

LlamaIndex (Python)

from Colnma import ColnmaClient
from llama_index.llms.openai import OpenAI
from llama_index.core.llms import ChatMessage, MessageRole

ptv = ColnmaClient(project_key="pv_sk_abc123")
system_prompt = ptv.get_prompt("hr-assistant", {"question": "What is our PTO policy?"})

llm = OpenAI(model="gpt-4o-mini")
resp = llm.chat(messages=[
    ChatMessage(role=MessageRole.SYSTEM, content=system_prompt),
    ChatMessage(role=MessageRole.USER, content="What is our PTO policy?")
])
print(resp.message.content)

HTTP → OpenAI (Python)

import requests
from openai import OpenAI

url = "https://api.colnma.com/api/sdk/v1/prompt/client/pv_sk_abc123/hr-assistant"
headers = {"Authorization": "Bearer YOUR_TOKEN", "Content-Type": "application/json"}
body = {"variables": {"question": "What is our PTO policy?"}}

system_prompt = requests.post(url, json=body, headers=headers).json()

client = OpenAI()
chat = client.chat.completions.create(
  model="gpt-4o-mini",
  messages=[
    {"role": "system", "content": system_prompt},
    {"role": "user", "content": "What is our PTO policy?"}
  ]
)
print(chat.choices[0].message.content)

Next Steps

Faq

What does the API return? The compiled system prompt with context.

Can I call prompts without variables? Yes — send

{}
or
{"variables":{}}

Documentation

Colnma Introduction

Menu

Introduction

Quickstart

SDKs

API

Guides

Next