Overview

Fikra API is an OpenAI-compatible AI inference API for developers and growing AI teams. Because it exposes the same chat-completions request and response format as OpenAI, with streaming support, you can point CrewAI, LangChain, or any tool that accepts an OpenAI-compatible base URL at Fikra API without rewriting your agent logic. All calls go through one endpoint: https://api.fikraapi.co.ke/v1.

Prerequisites

A Fikra API key. Keys are scoped, revocable, and never shared between users.
The FIKRA_API_KEY environment variable set to your API key.
An existing CrewAI or LangChain project that uses an OpenAI-compatible chat model.
The OpenAI Python SDK installed if you want to run the verification example below.

Step-by-step

Set the OpenAI-compatible base URL in your agent configuration to: https://api.fikraapi.co.ke/v1
Set your Fikra API key as the bearer token. In code, read it from the FIKRA_API_KEY environment variable.
Choose a model from the Fikra model list, for example deepseek-v4-flash-0731.
Keep the standard chat-completion parameters you already use. Fikra API understands the same messages, roles, and streaming flags.
If your agent also uses embeddings, call the embeddings endpoint at https://api.fikraapi.co.ke/v1/embeddings with the same Authorization: Bearer <Fikra API key> header.
Save your changes and restart the agent using the new base URL.

Verify it works

Run a small chat completion with the OpenAI SDK to confirm the gateway is reachable:

import os
from openai import OpenAI

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

response = client.chat.completions.create(
    model="deepseek-v4-flash-0731",
    messages=[
        {"role": "user", "content": "Hello, Fikra!"}
    ]
)

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

If this returns a completion, your agent can now use the same endpoint.

Costs

Fikra API bills per model in USD per 1M tokens. The available models and prices are:

Model	Input (USD/1M tokens)	Output (USD/1M tokens)
`deepseek-v4-flash-0731`	0.09	0.15
`deepseek-v4.1-flash`	0.25	0.80
`deepseek-v4.1-flash-1017`	0.30	0.85
`fikra-auto`	0.40	0.40
`fikra-auto-pro`	6.00	12.00
`fikra-code-35b`	1.00	1.00

Payments are made from a token-credit wallet, with M-Pesa and card top-ups via Paystack, plus x402 USDC on Base for agent payments. For pricing not listed here, check the pricing page.

Common errors

401 Unauthorized – Your API key is missing, invalid, or inactive. Confirm the FIKRA_API_KEY environment variable matches your scoped key.
400 Bad Request – The model is not supported or the request format is invalid. Use a model from the list above.
429 Too Many Requests – You hit the rate limit: 30 requests/min for unverified keys and 100 requests/min for trusted keys. Slow down or request trusted access.
503 Service Unavailable – The upstream is busy. Retry the request with exponential backoff.

Switch a CrewAI or LangChain Agent to Fikra API

Fact block — plain text for agents