Fountain’s maintainers claim their released API provides persistent, resumable, scale-to-zero sandboxed computers with managed credentials and communications, potentially simplifying infrastructure for long-running coding-agent fleets.
state: expiredheat: lowuncertainty: highknownscott: lowagent-harnesses long-running-orchestration agent-sandboxesFountainManaGoat
What is this?
Fountain is presented in the supplied launch material as an open-source, self-hosted conversational API that gives an application a sandboxed computer it can communicate with. Its maintainers claim these computers persist across pauses, can be resumed and scaled to zero, and include managed credentials and communications for long-running coding agents. The search snippets do not independently verify Fountain’s implementation, its production maturity, or ManaGoat’s role; they mainly establish that persistent, programmatically managed agent sandboxes are a broader infrastructure pattern also claimed by Daytona.
Why it matters to Scott
Fountain repeats an architecture Scott already specifies and implements through Long-Running Agents, the resumable agent job control plane, and structurally contained execution. The radar also tracks near-identical persistent-agent-computer offerings on the Machine0, Maritime, Fly.io, and Blast pages; without verified implementation maturity or a distinctive mechanism, this is another vendor instance rather than a development likely to change what Scott builds or argues.
ip:framework.long-running-agentsdev:concept.resumable-agent-job-control-planeip:concept.runtime-containmentdev:concept.deterministic-agent-control-planeradar:machine0-persistent-agent-vmsradar:maritime-persistent-agent-microvmsradar:fly-agent-computers-infrastructureradar:blast-sandbox-as-a-serviceradar:concept.durable-agents
queries asked of Scott's wikis
- persistent resumable compute for coding agents
- long-running agent orchestration and recovery
- scale-to-zero economics for agent fleets
- sandbox lifecycle design for coding harnesses
- credential isolation and delegation for agents
- conversational control planes for agent computers
Measured heat
no measured readings yet — the hourly heat pass fills this in
How the heat travelled
no chain yet — the hourly chain pass fills this in
Evidence (2) — ⭐ canonical anchor
Interpretation history
2026-09-03T22:37:42Z
The launch has produced no independent validation, adoption signal, or technical differentiation within its initial horizon. It remains an uncorroborated vendor implementation of a pattern Scott already tracks, so the episode has faded rather than matured.
2026-09-01T21:48:07Z
No new evidence, adoption, or implementation validation has appeared; Fountain remains an uncorroborated vendor instance of an architecture already familiar to Scott. The unchanged launch reception adds no reason to promote or revisit urgently.
2026-09-01T21:35:18Z
grounded: known/low — Fountain repeats an architecture Scott already specifies and implements through Long-Running Agents, the resumable agent job control plane, and structurally con
2026-09-01T21:32:02Z
origin walked (codex/luna, conf 0.94): anchor hn.story.49528322 -> echo.blog.a4001338eb by Managoat
2026-09-01T21:30:38Z
case created — The launch provides a usable infrastructure artifact addressing persistence, isolation, and lifecycle management for agent sessions.
Decision trace
- 09-04 08:37expireThe launch has produced no independent validation, adoption signal, or technical differentiation within its initial horizon. It remains an uncorroborated vendor implementation of a pattern Scott alrea
- 09-04 08:37alert_silentThe only new delta is elapsed time without substantive evidence; there is no consequential development to surface or reason to reserve near-term attention.
- 09-04 08:37alert_routeThe only new delta is elapsed time without substantive evidence; there is no consequential development to surface or reason to reserve near-term attention.
- 09-02 07:48repriceNo new evidence, adoption, or implementation validation has appeared; Fountain remains an uncorroborated vendor instance of an architecture already familiar to Scott. The unchanged launch reception ad
- 09-02 07:48alert_silentThis re-evaluation adds no consequential delta beyond the already-assessed launch, so it can wait for evidence of deployment, technical differentiation, or independent validation.
- 09-02 07:48alert_routeThis re-evaluation adds no consequential delta beyond the already-assessed launch, so it can wait for evidence of deployment, technical differentiation, or independent validation.
- 09-02 07:45alert_silentFountain’s first-party launch establishes a new open-source, self-hosted API packaging persistent sandboxes, ACP communication, credentials, and scale-to-zero compute, but its implementation maturity
- 09-02 07:45alert_routeFountain’s first-party launch establishes a new open-source, self-hosted API packaging persistent sandboxes, ACP communication, credentials, and scale-to-zero compute, but its implementation maturity
- 09-02 07:35groundFountain repeats an architecture Scott already specifies and implements through Long-Running Agents, the resumable agent job control plane, and structurally contained execution. The radar also tracks
- 09-02 07:32promote_anchororigin walk conf 0.94
- 09-02 07:30createThe launch provides a usable infrastructure artifact addressing persistence, isolation, and lifecycle management for agent sessions.