Summary:
Requesty raised $3 million in a seed round led by 20VC to act as a gateway between developers and AI providers like OpenAI and Anthropic.
The startup addresses cost and security risks in AI usage by centralizing controls, using prompt caching, and optimizing model routing.
Since pivoting to AI infrastructure, Requesty has attracted over 25,000 developers and achieved $1.5 million in annual recurring revenue.
Cofounders bring experience from companies like Amplitude and TikTok, positioning Requesty as a Cloudflare for the AI era.
The funding will support hiring to grow the team from 5 to 15-20 employees, focusing on engineering and growth.
The Rise of AI Gateways: Requesty's Vision
Leaders across industries are pushing employees to boost their AI usage, but this comes with rising costs and risks, such as uploading sensitive information to various large language models (LLMs). Requesty aims to address these challenges by reducing expenses and mitigating AI-related dangers.
Originally a data analytics firm, London-based Requesty pivoted this year to focus on AI infrastructure. Its technology serves as a gateway between AI providers like OpenAI, Anthropic, and Grok, and developers building AI-powered applications for tasks like coding, travel booking, invoice scanning, and translation.
Thibault Jaigu, Requesty's cofounder, compares the company to Cloudflare for the AI era, positioning it as a central hub connecting different AI servers. A common issue for companies is that engineers use multiple AI models without centralized governance, leading to security gaps and uncontrolled spending.
"The admins — such as IT security or DevOps — lack insight into what engineers are using, why, and the potential risks," Jaigu explained. Risks include exposing customer data or failing to revoke access keys after an employee leaves.
Requesty centralizes controls like spending limits and API access across various AIs. It also optimizes performance by routing requests through alternative providers during outages and uses prompt caching to save costs by reusing frequent prompts.
In just four months post-pivot, Requesty's platform is used by over 25,000 developers at companies like Mozart AI and Blue Morpho, achieving $1.5 million in annual recurring revenue. This places it in the emerging category of AI gateway startups, alongside others like Humanloop, which was recently acqui-hired by Anthropic.
Requesty announced a $3 million seed round led by 20VC, with participation from Tapestry VC, Insiders Ventures, and Tiny Supercomputer. The funds will support hiring, with plans to grow from 5 to 15-20 employees next year, focusing on engineering and growth teams.
Insights from Requesty's Pitch Deck
Requesty's tagline: 'The developer's gateway to safe AI' – Emphasizing security and accessibility.
Cofounder expertise – With backgrounds at Amplitude, TikTok, and Palo Alto Networks, bringing valuable experience.
Current AI challenges – Highlighting issues like lack of unified governance, security gaps, and poor cost control.
ROI demonstration – For a company spending $1 million, Requesty claims savings through prompt caching, automated audits, and risk mitigation.
Four-step process – Centralize AI, add controls, implement observability, and ensure safety measures.
Typical workflow – Analyzing requests for validity and optimization before reaching AI providers.
Growth areas – Focus on control/governance, visibility/accountability, and flexibility/optimization.
Feature benefits – Detailed look at how Requesty's tools enhance security and efficiency.
Comments