Official announcements, product launches, company updates, partnerships, and other developments from Fikra.
Tokens-per-minute and tokens-per-day caps are gone for topped-up accounts. Request quotas (3,000 RPM, 250k RPD) remain in force.
Top-up accounts now run 3,000 requests/minute and 1B tokens/day, 20x the previous daily token ceiling, effective immediately. Free-tier allowances unchanged. Agent traffic rides x402 (USDC on Base) with per-call limits published in pricing.md.
The Nairobi-based OpenAI-compatible inference API publishes DeepSeek at $0.045 per 1M input and $0.09 per 1M output, below both the official DeepSeek API and every observed gateway floor. Rates are served from a verifiable registry; agents pay per-call with x402.
Fikra is now available through the UwenzAi LLM-Gateway, giving users access to models from a Kenyan inference provider alongside providers hosted outside the region.