tangle/faster-test-loop
Run the failing test file after each edit. Full suite once before finishing.
Run AI agents in cloud sandboxes, with model access and hosted trace analysis.
Tangle gives developers and teams infrastructure for AI agents: isolated sandboxes, an agent runtime, model routing, approved integrations, and observability. Build an agent, run it, and inspect its traces before changing its behavior.
Sandbox, Router, and Intelligence are live developer entry points. A hospitality guest-request assistant is the first packaged use case, in preview.
Hospitality guest requests are the first hosted assistant preview
This preview aims to turn a guest message into an assigned task and alert a person when work is overdue. It is being prepared with a design partner; no live guest conversations yet. It is a hospitality workflow preview.
Give every agent an isolated workspace.
Coding harnesses, tool-calling agents, long-running workflows: each gets a real computer in a Tangle Sandbox with files, a terminal, and browser access.
$ clauderunning$ codexrunning$ opencoderunning$ hermesrunning$ openclawrunning$ nanoclawrunning$ kimirunning$ pirunningTurn each run into a reviewable record.
Capture the run
Record model calls, tool use, timing, tokens, and cost so a result can be inspected after the agent finishes.
illustrative trace · checkout-agent38.2s
run · fix flaky checkout test38.2s
claude-sonnet-5 · plan2.1s
read · checkout.test.ts0.5s
pnpm test · full suite12.8s
edit · retry-policy.ts1.9s
pnpm test · full suite11.4s
edit · checkout.test.ts0.9s
pnpm test · full suite8.9s
Find what repeats
Analyze multiple runs to find recurring failures and expensive steps worth testing as candidate changes.
illustrative trace · checkout-agent38.2s
run · fix flaky checkout test38.2s
claude-sonnet-5 · plan2.1s
read · checkout.test.ts0.5s
pnpm test · full suite12.8s
edit · retry-policy.ts1.9s
pnpm test · full suite11.4s
edit · checkout.test.ts0.9s
pnpm test · full suite8.9s
Review changes where you work
Turn measured findings into a pull request, issue, skill proposal, or team summary. A person decides what ships.
TAN-312
Adopt the failing-test-first loop across all coding agents.
skills/testing.md
Proposed testing guidance is ready for review.
summary
Review the candidate and its evidence before rollout.
Infrastructure for each agent run.
The five layers are sandbox, runtime, router, integrations, and observability. Sandbox and Router have live developer entry points. Runtime executes agent profiles in sandboxes. Integrations connect approved external services through grants. Their source code is available. Neither has a public start link. Intelligence is the hosted dashboard for agent traces. Send traces through OpenTelemetry or a Tangle SDK, or run an agent in Sandbox to record its work there.
Tangle plans an on-prem infrastructure option for teams that must keep agents and data in their own environment. An installation path is not available today.
For teams running agents
Bring a workload.
Bring representative tasks, an acceptance rule, permitted test data, and baseline bills.
We agree on scope, security requirements, limits, and trial charges before starting.
We run the agreed workload on Tangle, and on E2B where you permit it. The readout includes traces, accepted-task counts, and full-bill cost per accepted task after reconciliation. Failed attempts and retries count toward cost.
Apply by October 2 for an agreed readout target by October 19, 2026.
Tell us about your workloadStart with a free account.
The Free plan has no included credit. Add prepaid credit before using AI models. Sign in to compare paid plans, included usage, and limits before checkout.