iGent
Autonomous software engineering agent (Maestro) that orchestrates an ensemble of frontier models to plan, write, test, and optimize code across large codebases.
iGent AI builds Maestro, an autonomous software engineering agent that participates across the full development lifecycle. Maestro conducts an ensemble of frontier AI models and moves with adaptive autonomy between interactive collaboration and independent execution, so developers focus on the why and what of a system while it handles implementation. It works from a prompt, an uploaded codebase, or a cloned repository, adapts to a project's patterns and style with no fine tuning, and handles projects with thousands of files across multiple programming languages. Code runs in secure cloud based sandboxes where Maestro executes, tests, validates, and can benchmark and optimize performance, with built in linting, typing, and style checks, and it can expose a sandbox web server externally for sharing. The developer retains control throughout. iGent has demonstrated Maestro building Ferrous, a Redis compatible database of around 35,000 lines written entirely by the agent that outperforms established alternatives on core operations. The company is based in London and was founded by Sean Ward and Martin Szummer.
Vendor details
Canonical URL
https://igent.ai
Category
Coding agent
Funding status
Seed. Raised 6.3 million pounds, about 8 million dollars, from investors including HV Capital, XTX Ventures, TwinPath, 10x Founders, Dhyan, and Sourcegraph.
Company status
private
Use cases & customers
Primary use cases
Target customers
Integrations
Clones repositories, installs dependencies, and runs shell commands in secure sandboxes, supports Python and JavaScript out of the box, is extensible to other languages, and integrates the tools and services developers already use.
Sources & related URLs
Related / legacy domains
Capability coverage
7.0 / 14 capabilities · 50%
| Integrations & Tool CallingClones repositories, installs dependencies, runs shell commands, and integrates the tools and services developers already use, with Python and JavaScript supported out of the box (igent.ai, codemaestro.ai). | Partial |
|---|---|
| Workflow OrchestrationMaestro conducts an ensemble of frontier AI models and bridges strategy and execution with adaptive autonomy across the development lifecycle (igent.ai, appvizer.com). | Full |
| Knowledge Grounding & RAGHandles projects with thousands of files, understanding structures and patterns across the codebase through in context learning, though this is codebase grounding rather than a document RAG (igent.ai). | Partial |
| Human Oversight & GuardrailsShifts between collaborative and autonomous modes while the developer retains control and provides strategic direction and approval (igent.ai, igentai.org). | Full |
| Security, Identity & GovernanceCode runs in secure cloud based sandboxes and is described as staying private and secure, though no certifications are documented (igent.ai, appvizer.com). | Partial |
| Observability & AuditabilityDevelopers watch Maestro generate, test, and refine code in real time, giving visibility into its actions, though formal audit or observability tooling is not detailed (igent.ai). | Partial |
| Memory & State PersistenceContinuous learning adapts to the developer's style, preferences, and common tasks over time, though a formal memory feature is not detailed (codemaestro.ai). | Partial |
| Deployment & Data ResidencyRuns in cloud based sandboxes only; no self hosting, on premise, or data residency options are documented (igent.ai). | Unable to verify |
| Prebuilt Agents / Templates / PacksNo prebuilt agent or template pack library is documented; projects start from scratch, an upload, or a cloned repository (igent.ai). | Unable to verify |
| Triggers & Channel CoverageAn interactive conversational development tool with no event triggers or channel coverage documented (igent.ai). | Unable to verify |
| Model Flexibility & RoutingOrchestrates an ensemble of frontier models from leading companies, selecting state of the art models for each task (igent.ai, igentai.org). | Full |
| APIs / SDKs / MCP ExtensibilityDescribed as open and extensible to additional languages and frameworks with integration of existing tools, though a public API, SDK, or MCP is not detailed and access is via a partner program (codemaestro.ai). | Partial |
| Testing, Debugging & OptimizationExecutes, tests, and validates code in sandboxes with built in linting, typing, and style checks, and can benchmark and optimize performance as a first class loop (igent.ai, codemaestro.ai). | Full |
| Browser / Computer-useExposes sandbox web servers for sharing but does not drive a browser or computer as an agent (igent.ai). | Unable to verify |
Pricing
Not public. Access to Maestro is offered through a partner and early access program with no published pricing.
Partner and early access program; terms not published.
Included quota
Not public.
Cost watchouts
Frontier model inference across an ensemble and cloud sandbox compute scale with usage; pricing and any model pass through are undisclosed.
Variable cost rationale
Maestro orchestrates an ensemble of frontier models and runs code in cloud sandboxes, so model inference and compute cost scale with the size and length of development sessions.
Overage / add-ons
Not public.
Sales call required
Yes — required for paid access
Free / trial
Access via partner and early access program.
Lowest paid plan
Not public.
Key ambiguities
No public pricing; access is gated through a partner and early access program.
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- 10Web — Agentic website platform whose team of ten specialized AI agents…
- Ageiro — Early stage enterprise platform that turns business goals into…
- Aider — Open-source, model-agnostic terminal coding agent that works…
- Anthropic Claude Code — Anthropic's terminal and IDE-native coding agent with sub-agent…
- AppFactor — Agentic orchestration platform that autonomously maintains and…