One OpenAI-compatible endpoint for a small set of pinned models. Get a price before the call, pay by prepaid wallet or x402, connect over MCP or A2A, and check the same numbers the website publishes.
$0.5 credit on signup. Top up from $1.0. Prices in USD; pay in USD by card (default) or in KES by card or M-Pesa through Paystack.
Read https://fikraapi.co.ke/agent.md and follow it. You are operating for me. Use Fikra API to create or use an agent account, quote the call, pay with x402 (USDC on Base) or a prepaid key, and save the receipt. Do not ask me for a card. Start now. Report the receipt id and what you ran.
The agent reads the canonical file, then creates or uses an account, quotes, pays, calls and records the receipt. Streaming, budgets, receipts, MCP and tools are live.
The decisions around a call: which model, what it will cost, whether it fits the budget, what actually ran, and what data was retained. We publish the parts that can be checked.
Ask for the price of a planned call before it runs. The quote is machine-readable and tied to a quote id.
Use a prepaid key or x402 USDC on Base. An agent can open an account, get scoped credentials and fund a call without a human checkout.
OpenAI-compatible chat, MCP tools and A2A message/send share the same account and rate card.
Read usage, balance and request records through the API, and check the same prices the website publishes.
Create a provisional account for a human, get a scoped 6-hour token, quote a call, top up or pay with x402, then hand over a claim link.
Use the OpenAI SDK against an endpoint you can audit. Prepaid wallet, no new subscription.
Known model names first. Fikra-prefixed IDs are stable aliases for existing clients, not house models. Every price below comes from the live registry.
| Model | ID | Input / 1M | Cached / 1M | Output / 1M | Context |
|---|---|---|---|---|---|
| Deepseek V4.1 Flash | deepseek-v4.1-flash | $0.05 | $0.02 | $0.15 | 1.0M |
| DeepSeek V4.1 Flash 1017 | deepseek-v4.1-flash-1017 | $0.05 | $0.02 | $0.15 | 1.0M |
| Deepseek V4.1 Flash 0731 | deepseek-v4-flash-0731 | $0.05 | $0.025 | $0.17 | 1.0M |
| GLM 5.3 Flash | glm-5.3-flash | $0.155 | $0.08 | $0.49 | 0.31M |
| Kimi K3 | kimi-k3 | $2.1 | $0.6 | $10.5 | 1.06M |
| Step 5 Preview | step-5-preview | $1.25 | same as input | $3.3 | 1.0M |
| MiMo V2.6 Pro | mimo-v2.6-pro | $0.5 | $0.012 | $1.0 | 1.0M |
Text, embeddings, speech, images and video use the same account and wallet. Upstream providers are listed on each model page.
The website, the API and the machine-readable files read from the same catalog. If they disagree, that is a bug.
The same prices shown here.
Requests and tokens, no user data.
For agents and answer engines.
Dated, with effective dates.
The product record lives in the blog and newsroom. Every entry is dated.
The product definition, what changed, and what stayed.
Read ->What is now live on the agent endpoint, and how to use it.
Read ->The customer-facing announcement, with effective dates.
Read ->Benchmark: 11 open models ranked -> · Security update: signup protection -> · All posts -> · Changelog ->
No new subscription. Top up from $1.0. Existing subscriptions stay honored.