Jetty
Also known as: Jetty AI, Jetty Solutions
Infrastructure for running agentic AI workflows: one OpenAI-compatible endpoint that proxies over 100 LLM providers, runs autonomous agents in isolated sandboxes from markdown runbooks, and applies evaluation loops with full trajectory tracing.
Jetty (Jetty Solutions Inc, Montreal) builds an infrastructure layer for reliable agentic AI. A single OpenAI-compatible chat-completions endpoint operates in two modes: a passthrough proxy across more than 100 LLM providers that records every call as a trajectory, and a runbook mode that provisions an isolated sandbox, runs an autonomous agent defined by a markdown runbook, and returns structured results. Agents run end-to-end workflows, evaluate their own output against rubric or programmatic checks, and retry within bounded iteration until they meet a defined quality bar, with a human in the loop. Trajectories capture full execution traces for observability, and a step library of more than 47 prebuilt activities covers AI models, control flow, data processing, and evaluation. The company raised more than 2 million dollars in pre-seed funding led by AQC Capital and Layers Capital, with participation from Mila and strategic angels.
Vendor details
Canonical URL
https://jetty.io
Category
Agent infrastructure
Funding status
Private, pre-seed. Raised more than 2 million dollars led by AQC Capital and Layers Capital, with participation from Mila Ventures and strategic angel investors. Founded in Montreal, Canada; CEO and founder Jonathan Lebensold.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
OpenAI-compatible single API endpoint that works with any OpenAI SDK and proxies more than 100 LLM providers; a step library of more than 47 prebuilt activities covers AI models, control flow, data processing, and evaluation. Agents execute tools inside isolated sandboxes.
Sources & related URLs
Related / legacy domains
Capability coverage
7.5 / 14 capabilities · 54%
| Integrations & Tool Callingjetty.io and docs.jetty.io: single endpoint runs agents that call tools in an isolated sandbox plus a step library of 47-plus prebuilt activities; no broad third-party connector catalog documented. | Partial |
|---|---|
| Workflow Orchestrationdocs.jetty.io: orchestrates multi-step evaluation pipelines and runs end-to-end agent workflows defined by markdown runbooks with control flow. | Full |
| Knowledge Grounding & RAGNo knowledge grounding or RAG capability documented on jetty.io or docs.jetty.io. | Unable to verify |
| Human Oversight & Guardrailsjetty.io company page: agents iterate with a human in the loop until they meet a defined quality bar. | Full |
| Security, Identity & Governancejetty.io: isolated sandbox execution and audit trails; no SSO, RBAC, or compliance certification documented. | Partial |
| Observability & Auditabilitydocs.jetty.io: trajectories capture full execution traces with inputs, outputs, and metadata for observability. | Full |
| Memory & State PersistenceNo first-class agent memory documented; snapshot environments provide reproducibility, not memory. | Unable to verify |
| Deployment & Data ResidencyHosted API (flows-api.jetty.io) only; no self-host, VPC, or options documented. | Unable to verify |
| Prebuilt Agents / Templates / Packsdocs.jetty.io: step library of 47-plus prebuilt activities; no prebuilt end-user agents. | Partial |
| Triggers & Channel CoverageInvoked via API calls only; no scheduled or event-based triggers documented. | Unable to verify |
| Model Flexibility & Routingdocs.jetty.io: proxies 100-plus LLM providers; switch providers by changing the model field. | Full |
| APIs / SDKs / MCP Extensibilitydocs.jetty.io: one OpenAI-compatible API that works with any OpenAI SDK out of the box. | Full |
| Testing, Debugging & Optimizationjetty.io: managed eval batches with rubric and programmatic validation and bounded retry loops that hill-climb to a quality bar. | Full |
| Browser / Computer-useNo browser or computer-use capability documented. | Unable to verify |
Pricing
Not publicly listed
Variable cost rationale
A usage-based LLM-proxy and sandbox-execution API implies per-call and per-token cost, but no rate card was published this session.
Sales call required
Mixed (some tiers require a call)
Free / trial
Not documented
Related vendors
- Acrab — Singapore compute infrastructure company building a full stack…
- AgentOps — Agent observability and reliability platform with broad model and…
- Agno — High-performance agent runtime and framework (formerly Phidata) with…
- AIsa — Unified resource and payment gateway for AI agents that lets them…
- AlphaBitCore — AI control plane that governs how models, agents, tools, and…
- Anchor Browser — Cloud hosted browser infrastructure that lets AI agents operate real…