About Fleece AI
We build AI agents that actually do the work
What started as an AI agency in 2023 became Fleece AI App in late 2025 — a delegative workspace where autonomous agents execute workflows 24/7 across 3,000+ app integrations. Today we offer a hybrid model: the App for everyday automation, and our agency for complex custom projects that need a human team behind the scenes.
Our mission
We believe every team deserves an AI workforce that works as hard as they do. Fleece AI eliminates the gap between strategy and execution — you tell your agents what needs to happen, and they connect the right apps, run the workflows, and handle edge cases without you watching over them.
The platform
Fleece AI is built on three pillars: autonomous agents that reason and execute multi-step tasks, deep integrations via Pipedream connecting to 3,000+ apps (Gmail, Slack, Notion, Salesforce, HubSpot, GitHub, Stripe, and more), and a natural language interface that lets you describe any workflow without writing code.
Our team
We are a Paris-based team of engineers and AI researchers. Our roots go back to 2023 when we started as an AI agency — that experience building real projects for real clients shaped how we built Fleece AI App. We combine the speed of our platform with the depth of our agency expertise: start with the App, and when projects get complex, our agency team steps in. We ship frequently — our changelog shows every user-facing change, dated. We take security seriously: all data is encrypted at rest and in transit, we maintain immutable audit logs, and SOC 2 Type II certification is in progress.
Technology
Our platform runs on Next.js 15, React 19, and Drizzle ORM on Neon Postgres. Agents use leading models — GPT-5.2, GPT-5.6 (Luna, Terra, Sol), Claude Opus 4.7/4.8, Claude Fable 5, and Mistral Medium 3.5 — with a dynamic step budget that scales based on your connected apps. External integrations reach the LLM through a remote MCP server with a 128-tool cap and priority trimming.
Get in touch
Have a question about our platform, want to discuss an enterprise deployment, or simply want to say hello? Reach us at contact@fleeceai.app or through our contact page.
Contact usCommon questions
Where is Fleece AI based?
Fleece AI is headquartered in Paris, France. Our team is distributed across Europe, and we serve customers worldwide. Our platform supports English and French.
What models do your agents use?
Fleece AI supports GPT-5.2, GPT-5.6 (Luna, Terra, Sol), Claude Opus 4.7/4.8, Claude Fable 5, and Mistral Medium 3.5. Model access is gated by your plan tier — GPT-5.2, GPT-5.6 Luna, and Mistral Medium 3.5 on every plan, GPT-5.6 Terra on Pro, and Claude Opus 4.8 / GPT-5.6 Sol on Business.
How does Fleece AI handle data privacy?
All data is encrypted at rest and in transit. We maintain immutable audit logs for all agent actions. We offer a GDPR-compliant Data Processing Agreement, and SOC 2 Type II certification is in progress. We do not sell your data or use it to train models.
Can I use Fleece AI without coding?
Yes. Fleece AI is designed to be code-free. You describe workflows in natural language, and our agents handle the execution across connected apps. If you do want to go deeper, our REST API and OpenAPI spec let you integrate programmatically.