Led · Seed · Apr 2026

Maxed provides an AI-native operating system for CPA firms built around two agentic AIs—Max, which runs and monitors back-office workflows, and Ed, a client-facing agent that engages clients through a firm-branded portal while keeping accountants in control. The platform is designed to automate repetitive administrative work so firms can serve more clients without adding headcount and recover billable hours lost to manual tasks. Maxed is currently in beta, working with early design partners and accepting requests for early access. The company is led by founder and CEO Fifi Siddiqui and is headquartered in Miami. Public plans include expanding product features and driving broader industry adoption using proceeds from the recent pre-seed round.

Led · Seed · Oct 2025

Flamingo is a Miami-based startup that has built OpenFrame, a platform combining autonomous AI agents with open-source infrastructure to overhaul the economics of the $380 billion MSP market. Its two agents, Fae and Mingo, autonomously handle client-side tasks such as password resets and system patches, as well as back-end operations like threat detection and routine maintenance, all while keeping technicians in the approval loop for sensitive actions. The product is offered either self-hosted at no cost or as a managed service with per-device pricing, bundling RMM, remote access, MDM, and an integrated SIEM. More than 1,000 MSPs—about 3 % of the market—have joined the waitlist, and 150+ early adopters are actively testing the software in Flamingo’s Slack community. By removing vendor lock-in and slashing labor costs, the company claims it can help MSPs improve net margins from the industry’s typical 8–12 % to as high as 50–60 % EBITDA. Backed by $2.2 million in pre-seed funding, Flamingo plans to expand its sales, support, and engineering teams to meet growing demand and continue refining its community-driven roadmap.

Participated · Seed · May 2025

Coworker.ai offers an autonomous AI agent—marketed as a "Manus for Companies"—that can research, plan, and execute multi-step tasks by integrating with 25+ enterprise tools like Jira, Slack, GitHub, and Salesforce. Its core technology is OM1, a proprietary organizational memory architecture that tracks 120+ business dimensions (projects, teams, meetings, documents) to retain accurate, company-specific context. The agent is positioned to act as a reliable teammate across sales, engineering, customer success, and leadership workflows, automating tasks such as proposal drafting, code writing, release notes, and health-risk flags. The product launch emphasizes security, privacy, source accuracy, and permissions, citing compliance with SOC 2, GDPR, and CASA Tier 2. Coworker.ai cites customer impact with a testimonial from Curri’s CPO and examples of automated sprint recovery. The company was founded by former Uber executives Alex Calder and Bradford Church and announced its launch from San Francisco.

Participated · Seed · Jan 2025

Thoras builds an AI-driven reliability and observability platform that helps engineers rapidly locate root causes of software disruptions and predict demand fluctuations. The company emphasizes cost-efficient reliability, using smaller machine-learning models rather than large LLMs to avoid overconsuming cloud resources. Thoras claims it can find and solve issues 70% faster and save customers up to 60% on cloud costs. It has been focused on Kubernetes environments with plans to expand into other types of cloud software. Thoras came out of stealth in January 2024 and reports revenue growth of 360% in the past nine months. The startup intends to use new financing to hire engineers, build out product capabilities and keep up with customer demand. Thoras.ai builds software that monitors cloud services to forecast demand and autonomously scale Kubernetes workloads up or down. The agent integrates alongside a service and connects directly to development environments with no APIs and no data leaving the customer environment, prioritizing security and privacy. Developers get a dashboard that surfaces resource metrics and performance alerts, with an emphasis on a visually appealing UX. The company currently uses task-based machine learning to forecast consumption and proactively identify problems, and it plans to add LLM-powered NLP features in future roadmap products for troubleshooting. The founders launched the company right after the first of the year, released an initial product version, are running in live customer environments, and are already generating revenue. Today they announced a $1.5M pre-seed financing.

Led · Seed · Dec 2024

On Me, founded by former Google employees, has built a digital gifting platform that replaces plastic gift cards with flexible, interest-based digital cards. Users can choose among 72 categories—such as running, reading, gaming, or concerts—and attach personalized video messages, photos, or GIFs before sending. The cards work across iOS, Android, and web and can be stored in Apple Pay or Google Wallet, reducing plastic waste in a market that produces 30 billion cards annually. To broaden its appeal, the company plans to add experiential categories like horseback riding lessons, wine tastings, and theme park trips. Since launch, On Me has facilitated more than $2.5 million in gifts for over 26,000 users and is growing about 50 percent month-over-month. The global gift card market is expected to reach $2.3 trillion by 2030, positioning On Me for significant expansion. Financially, the company had previously secured a $1.7 million pre-seed round and has now added a $6 million seed round to accelerate product and category growth.