Azumo opens early access to Valkyrie, a flat-rate, open-weight coding agent service, starting with a friends and family test period ahead of full public release
Azumo, a nearshore software development company specializing in AI-powered applications, today announced the launch of Valkyrie, an open-weight coding agent service designed to give engineering teams a flat, predictable cost for AI-assisted development. Valkyrie is opening first to a small group of friends and family for an initial period, ahead of a full public release.
Engineering teams hit a wall with metered AI billing. Valkyrie removes it: flat per-seat pricing, no counting tokens.”
— Chike Agbai, Founder & CEO of Azumo
Valkyrie runs open-weight models, including Qwen, Llama, DeepSeek, Mistral, and GLM, behind a compatible API that plugs directly into the coding tools developers already use, including Claude Code, Cursor, or their command line. Azumo operates the infrastructure underneath, including provisioning, scaling, and reliability, so engineering teams get agentic coding support without standing up or managing GPU infrastructure themselves.
Unlike metered API billing, Valkyrie is priced per developer seat, with no usage ceiling. Every change produced through Valkyrie can be paired with Azumo Code Audit, an automated review layer that scans code for security, cost, and architecture issues and returns a prioritized fix list before changes reach production.
Marketing Technology News: MarTech Interview with Theresa Pham, Head of Product @ Wayvia
“Every engineering team we talk to is running into the same wall: AI coding agents are becoming essential, but the metered bill makes it hard to let your best developers use them as much as the work actually requires,” said Chike Agbai, Founder and CEO of Azumo. “Valkyrie removes that ceiling. Teams get a flat cost per seat and the freedom to build without counting tokens.”
The initial friends and family phase lets a limited set of teams run Valkyrie in real development workflows before Azumo opens broader access. Azumo says the initial phase will focus on Builder Teams running proofs-of-concept and early production jobs, with support for regulated and higher-throughput Scale and Enterprise workloads to follow.
“We built Valkyrie to run alongside the tools teams already use”, said JP Lorandi, Chief Technology Officer of Azumo. “The friends and family phase is where we stress-test that in real workflows, watching how teams actually mix open-weight and commercial models on a single codebase, before we open it up more broadly.”
Marketing Technology News: Idle data is as good as no data
Valkyrie builds on years of AI infrastructure and applied AI work at Azumo, documented across the company’s AI case studies.
Valkyrie is built on Azumo’s SOC 2 certified infrastructure, with GDPR-aligned data handling, role-based access control, encrypted communications, and detailed audit logs. Customer code and prompts are used only to run each request and are never used to train a model.
