ST. PETERSBURG, Fla., April 21, 2026 (GLOBE NEWSWIRE) — Osirus AI today announced the public launch of its enterprise AI platform, built from the ground up by three founders with no outside funding. The platform combines every major AI capability — Chat, Search, Image, Video, Speech, Storage, and a model marketplace — into a single workspace. Osirus then layers its purpose-built Agent Studio on top for organizations to create and operate AI agents at scale. Alongside this launch, Osirus is releasing a limited preview of Codex Bridge, its multi-model AI coding extension for Visual Studio Code.
The announcement comes as every major AI platform — from Anthropic to OpenAI to Perplexity — shifts its focus from general AI assistants toward agents. Osirus was built for exactly this moment: a platform where every tool needed to create, train, deploy, and improve AI agents lives in one place, connected to your team, your data, and your choice of AI provider.
“AI isn’t just about models anymore – it’s about the agents doing the work. Osirus is the system enterprises need to build, deploy, and manage agents with visibility and control so that they can scale them across their business.”
— Shawn Moore, Co-Founder & CTO, Osirus AI
The Problem Osirus Solves
Most teams today run several AI subscriptions with no shared history, no team permissions, and no way to track what was created or by whom. Additionally, building AI agents has required engineering resources and deep integrations with individual providers, locking organizations into a single model or cloud platform even as better options emerge.
Osirus solves both problems. It gives every team member a single place to do all their AI work, with shared projects and role-based access. And it gives organizations the infrastructure to build agents that run on any AI provider that fits their budget and compliance requirements — without rebuilding when the models change.
Introducing the Osirus AI Platform
The Osirus AI platform gives teams a unified workspace with Chat, Search, Image, Video, Speech, and Storage — all connected to a shared project system with user permissions and full history. Every AI interaction and asset lives in one place your team can actually manage, instead of scattered across multiple subscriptions.
The platform connects to every major AI provider through a built-in Provider and Model Marketplace, where AWS Bedrock, Google Vertex, Microsoft Azure, OpenAI, Anthropic, and Hugging Face are all supported. Frontier models — including Amazon Nova Pro, Claude Sonnet, and GPT-4o — are also available and can be switched on the fly while maintaining full context. Organizations use their own cloud accounts, so data stays within their existing governance infrastructure.
A free Developer account with test tokens is available at osirus.ai with individual upgrades via the Osirus Pro and Max plans. Enterprise organizations can access the full platform with Agent Studio through Osirus Org and Team plans. Details at osirus.ai/pricing.
Announcing Today: Osirus Agent Studio
Also launching today is Osirus Agent Studio, a purpose-built environment for organizations that need to build, deploy, and continuously improve AI agents at scale.
With Osirus Agent Studio, AI doesn’t replace domain experts — it promotes them, transforming your staff into AI managers. The agent handles the front-line interaction while your team reviews every conversation, flags issues, and improves responses. The organization gets the throughput of AI with the accountability of human oversight.
In Osirus, agents are trained using a built-in website crawler that ingests content from any domain and any number of pages and uses it as the foundation for agent knowledge. Organizations can build department-specific agents trained on their own content, policies, and documentation — without writing code.
“The organizations getting real value from AI right now are not replacing their people — they are promoting them. You train an agent on your content, deploy it to the front line, and your team manages it like any other employee: reviewing its work, correcting mistakes, and making it better every day.”
— Dasha Moore, Co-Founder & COO, Osirus AI
Agent Help Desk: How Agents Improve Over Time
Every conversation an agent has generates a ticket in the Osirus Agent Help Desk. AI managers review these tickets, flag problems, and apply corrections using the built-in Prompt Template Management system. This is how agents improve continuously after deployment — not through periodic retraining by an AI team, but through daily input and context from the people who understand the work.
Running in Production Today
One of Osirus’s early enterprise customers is running 20 AI agents across departments, with 20 active users and website crawls across almost two dozen separate domains. Each agent is trained on department-specific content and managed by staff who review interactions and improve responses through the Agent Help Desk. The organization is actively expanding its deployment into new departments — without adding headcount.
Agent Studio Capabilities
- Website Crawler: Train agents on any web content, any domain, or any number of pages. No data pipeline or engineering work required.
- Agent Help Desk: Every conversation generates a reviewable ticket. Teams can flag, correct, and improve agent responses continuously.
- Prompt Template Management: Access a full library of prompt templates and reinforcement tools to shape agent behavior across every scenario.
- Multi-Provider Support: Deploy agents on AWS Bedrock, Google, Azure, OpenAI, Anthropic, or Hugging Face — using your organization’s own cloud account.
- Model Flexibility: Choose from Amazon Nova Pro, Nano Banna, Claude Sonnet, GPT-4o, and more. Assign the best model per agent based on performance and cost.
- Role-Based Access: Staff are assigned as AI managers with appropriate permissions per project, department, or agent.
Also Launching Today: Osirus AI SDK
The Osirus AI SDK lets you make any agent built with Agent Studio embeddable in any website or mobile application. The package is published on npm, and developers can install it for full programmatic control or drop it in as a single script tag. An agent trained on a department’s website can be live on its page in minutes, with no backend integration required.
- npm install: Full SDK for developers building custom integrations or embedding agents into existing web and mobile applications. Available now at npmjs.com/package/@osirus/agent.
- CDN / UMD script tag: One-line embed for any webpage — no build step, no npm required. Drop it in like Google Analytics.
- Mobile support: SDK supports integration into native mobile applications alongside web deployments.
Preview Release: Codex Bridge for VS Code
Available today in limited preview, Codex Bridge is an AI coding extension for Visual Studio Code that connects to every major AI provider and model — giving developers the freedom to choose the right model for the task without leaving their editor.
Coding Models Available in Codex Bridge
- Claude Sonnet 3.7: Anthropic’s latest model, strong on reasoning-heavy refactors and multi-file context.
- GPT-4o: OpenAI’s flagship model for fast, high-quality code generation and explanation.
- GPT-4.1: OpenAI’s latest coding-optimized release with improved instruction following.
- Gemini 2.5 Pro: Google’s top coding model with a 1M token context window for large codebases.
- Amazon Nova Pro: AWS Bedrock’s high-performance model, ideal for teams already running on AWS infrastructure.
Codex Bridge runs natively in VS Code with no browser tab or context switch required, and models can be switched mid-session. The source is available on GitHub at github.com/techcto/codex-bridge.
Codex Bridge is currently available to a select group of early access customers. Organizations interested in joining the preview can contact info@osirus.ai.
Seeking Strategic Investment and Partners
Osirus AI is actively seeking strategic investment to scale its enterprise agent platform and build a national sales organization. The company is bootstrapped and in production, with paying enterprise customers running AI agents today.
The market timing is not coincidental. ARK Invest’s Big Ideas 2026 report, published by ARK founder Cathie Wood — who operates from St. Petersburg, the same city where Osirus AI is headquartered — identifies AI agents as one of its central investment themes for the year, projecting capital investment in AI infrastructure to reach $1.4 trillion by 2030. ARK’s own recent moves reinforce the thesis: the firm committed approximately $240 million to OpenAI in March 2026 and has identified enterprise AI agent platforms as a key area of disruptive innovation.
Osirus AI is not building AI models. It is building the system enterprises use to deploy and manage agents built on top of those models — a layer that every major AI provider currently lacks and cannot easily build without alienating their own customers. The platform is provider-agnostic by design, which means it grows alongside the AI ecosystem rather than competing with it.
Why Osirus AI is Built to Scale
- Self-service deployment: An enterprise can have its first agent trained, deployed, and live on its website by the end of a single demo. No professional services, no long implementation cycle.
- Sticky by design: Agents are trained on an organization’s own content and managed by their own staff. The longer they run, the more valuable they become.
- No vendor lock-in: Osirus does not hold the keys to an organization’s AI. They use their own cloud accounts across AWS, Google, Azure, OpenAI, and Anthropic. This is a feature enterprise procurement teams specifically ask for.
- Scales from a coffee shop to a Fortune 500: The same platform that runs 20 agents for a government organization can onboard a small business with a single agent in an afternoon. The pricing model scales accordingly.
- National sales opportunity: Every enterprise, government agency, and mid-market organization is currently evaluating AI agents. Osirus has a working product, live customers, and a repeatable demo. A funded sales team can move immediately.
“The question every organization is asking right now is not whether to deploy AI agents. It is how to do it without getting locked into a single vendor, without putting their data at risk, and without betting the whole thing on one provider’s roadmap. That’s exactly the problem Osirus was built to solve.”
— Matt Garrepy, Co-Founder & CDO, Osirus AI
Osirus AI is open to conversations with strategic investors, enterprise partners, and acquirers who see the agent layer as a critical missing piece in their AI stack. Contact info@osirus.ai.
Our Approach to AI
Osirus AI was built on a simple principle: organizations should always be in control of their AI. With Osirus, every agent interaction is logged and reviewable. Role-based access controls determine who can build, manage, or inspect agent activity. Organizations run on their own cloud provider accounts, which means AI-generated data stays within their existing governance and compliance infrastructure.
The Agent Help Desk model reflects a deliberate stance: AI behavior should be visible and addressable by the people closest to the work. When an agent gives a poor response, that failure is documented and directly improvable, so organizations aren’t left managing a black box.
Supporting every major provider is also a choice, not just a feature. Different organizations have different compliance requirements, existing vendor relationships, and budget structures. Osirus doesn’t tell customers which AI to trust — it gives them a system that works with whatever they already use or decide to use next.
Availability
The Osirus AI Platform and Agent Studio are available now. A free Developer account with test tokens is available at osirus.ai.
Individual plans (Pro and Max) and Enterprise plans (Org and Team) are at osirus.ai/pricing.
The Osirus AI SDK is available via npm and CDN. Codex Bridge for VS Code is in limited preview — contact info@osirus.ai for early access.
About Osirus AI
Osirus AI is an enterprise AI platform that gives organizations a single workspace for all AI work — including chat, search, image generation, video, speech, and storage — alongside a full Agent Studio for building, deploying, and managing AI agents trained on organizational content. Osirus supports all major AI providers, including AWS Bedrock, Google Vertex, Microsoft Azure, OpenAI, Anthropic, and Hugging Face, and connects to frontier models including Amazon Nova Pro, Nano Banna, Claude Sonnet, and GPT-4o. Osirus is headquartered in St. Petersburg, Florida, with U.S.-based innovation and support.
For more information, visit osirus.ai.
Media inquiries: Matt Garrepy | info@osirus.ai | http://www.osirus.ai
