Amp
Amp is a standalone vertical AI agent from Amp (Amp, Inc.). A coding agent that runs in the terminal and on remote machines, working on threads on its own and waking on schedules or events. Nearfi's Agent Index rates its autonomy 4/5 (multi-step on its own): Amp's page says agents run unsupervised on remote machines called Orbs, keep working after the laptop is closed and wake on schedules or events such as a new Linear issue or failing CI. It stops short of a 5 because the user starts each thread and reviews the changes. Status: live; 0 incident(s) logged.
By Amp (Amp, Inc.) · launched 2025-05 · last checked 2026-11-05
Why 4: Amp's page says agents run unsupervised on remote machines called Orbs, keep working after the laptop is closed and wake on schedules or events such as a new Linear issue or failing CI. It stops short of a 5 because the user starts each thread and reviews the changes.
What it does. Amp runs from a CLI, which can also run headless on servers and CI, and from the web, with threads, previews of running apps and a tab for reviewing changes. Each thread gets a fresh remote machine with the code and tools. Sourcegraph launched Amp in 2025 and on 2 December 2025 announced it was becoming a separate company, Amp, Inc., which says it is profitable.
Where it falls short. The launch month (May 2025) comes from Tessl's report; Amp's own pages opened give no launch date. The profitability statement is the company's own and no figures are given.
What it can do
Incidents
- None logged.
Closest fiction
Silicon Valley — Engineering work delegated to a tireless junior.
Related coding agents
- Augment Cosmos Augment Code · autonomy 4/5
- Claude Code Anthropic · autonomy 4/5
- Codex OpenAI · autonomy 4/5
- Cursor Anysphere · autonomy 4/5
- Devin Cognition · autonomy 4/5
- Factory Droids Factory · autonomy 4/5
History
- 2026-11-05 Added to the index.
Sources
- Amp Amp
- Amp Inc Amp, 2025-12-02
- Sourcegraph spins out AI coding agent Amp as a standalone company Tessl, 2025-12-03
← All agents in the index · Autonomy is rated on what the product does today, on a 1–5 scale. Every fact links to a source.