Overview

Fikra API is an OpenAI-compatible inference API that lets you send chat-completion requests to run models behind a single https://api.fikraapi.co.ke/v1 endpoint. Because requests and responses follow the OpenAI chat-completions format, you can use standard OpenAI SDK tooling to process earnings-call transcripts and return a summary. This page shows how to build a morning script that reads a transcript, asks a model to summarize it, and prints the result.

Prerequisites

A Fikra API key. Set it as an environment variable named FIKRA_API_KEY.
Python with the openai SDK installed.
A transcript file you want to summarize (for example, transcript.txt).
The API base URL: https://api.fikraapi.co.ke/v1.
A model from your Fikra API account. The example below uses deepseek-v4-flash-0731.

Step-by-step

Configure the OpenAI client to point at Fikra API and read your key from the environment.

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.fikraapi.co.ke/v1",
    api_key=os.environ["FIKRA_API_KEY"],
)

Load the transcript and send it to the chat completions endpoint with a summarization instruction.

transcript = open("transcript.txt").read()

response = client.chat.completions.create(
    model="deepseek-v4-flash-0731",
    messages=[
        {"role": "system", "content": "You summarize earnings calls."},
        {"role": "user", "content": f"Summarize this earnings call:\n\n{transcript}"},
    ],
)

print(response.choices[0].message.content)

Run the script each morning using a scheduler such as cron. The script reads the latest transcript, calls POST https://api.fikraapi.co.ke/v1/chat/completions, and outputs the summary.

Verify it works

Run the script with a sample transcript. If the API key is valid and the model name is correct, you should see a plain-text summary printed to the console. The response is an OpenAI-style chat-completion object, so you can inspect response.choices[0].message.content for the summary text.

Costs

Pricing is per model. The example model deepseek-v4-flash-0731 costs USD 0.09 per 1M input tokens and USD 0.15 per 1M output tokens. Other models on the platform have different prices; check the pricing page before picking a model. Payments are handled through a token-credit wallet with M-Pesa and card top-ups via Paystack, and x402 USDC on Base for agent use.

Common errors

401 to The API key is missing, invalid, or inactive. Check that FIKRA_API_KEY is set and the key is still valid.
400 to The requested model is not supported. Verify the model name against the models available on your account.
429 to You hit the rate limit: 30 requests/min for unverified keys, 100 requests/min for trusted keys. Slow down or wait.
503 to The upstream service is busy. Retry the request with exponential backoff.

Summarize Earnings Calls Automatically Each Morning

Fact block, plain text for agents