Show HN: 기술적인 문제를 진단하기 위해 AI 에이전트를 구축했습니다.
hackernews
|
|
📰 뉴스
#claude
#오픈소스
원문 출처: hackernews · Genesis Park에서 요약 및 분석
요약
Rinhelp은 개발자가 지원 티켓 해결에 시간을 낭비하지 않도록 돕는 AI 에이전트로, Crisp의 지원 스레드에서 시작됩니다. 이 도구는 GitHub 코드 문맥과 Sentry 런타임 증거를 자동으로 수집하여 실제 증거에 기반한 진단 초안을 작성하고 제공합니다. 이를 통해 사용자는 별도의 조사 없이 초안을 검토하고 빠르게 문제를 해결하거나 수정할 수 있어 개발 작업 흐름을 유지할 수 있습니다.
본문
Five tools open before you can answer one technical issue. Diagnose issues w/o babysitting another agent Rinhelp starts from a Crisp support thread, checks GitHub code context and Sentry runtime evidence, then returns an evidence-backed diagnosis draft for human review. For founders and engineers trying to keep shipping Works with your stack Every technical ticket pulls you out of the work you meant to ship One technical ticket can pull you out of shipping into manual investigation, bouncing between support context and another agent you still have to steer. One bug report can blow up the rest of your build session. Generic agents still need you to frame and steer the investigation. You still have to judge whether the answer is backed by real evidence. Rinhelp handles the first investigation pass inside your support workflow Rinhelp starts where the issue already happens: the Crisp support thread. It pulls GitHub code context, Sentry runtime evidence, and available account context together, then hands back an internal draft you can review before you reply, triage, or fix anything. Starts from Crisp Rinhelp starts where the issue already happens instead of sending you into a second workflow. Checks GitHub and Sentry It checks the Crisp conversation, read-only GitHub repo context, Sentry runtime evidence, and available account context in one investigation pass. Stays internal until review You get a draft to review before any reply, triage decision, or product change goes out. From support thread to reviewable diagnosis draft Rinhelp runs the first pass, then hands back an internal draft for review instead of forcing you to piece the workflow together by hand. A technical issue comes in through Crisp A Crisp bug report gives Rinhelp the starting point for the investigation. GitHub and Sentry context get pulled together Rinhelp gathers the support thread, read-only repo context, runtime signals, release context, and account data available for that workspace. Evidence gets checked It investigates the issue and checks whether the evidence is strong enough to support a bounded diagnosis. You get a draft to review The output comes back as an internal draft you can review and use as the starting point for the next step. More time to keep shipping Faster first-pass diagnosis, less manual digging, and a clearer review loop before anything reaches the customer. Protect shipping momentum Let Rinhelp handle the first pass so one technical ticket does not blow up the rest of your build session. Evidence before guesses Rinhelp is built to support claims with real support, code, runtime, and data evidence instead of fluent speculation. A safer review loop You get a diagnosis draft to review before anything goes out to the customer or changes in the product. Honest uncertainty If the evidence is not strong enough, Rinhelp marks the review draft as blocked and shows the exact context to collect next. Try Rinhelp before you pick a plan Your first 5 investigations are on us. When you are ready to continue, choose the monthly workspace plan that matches your technical investigation volume. Starter For smaller teams - Up to 50 investigations - 1 workspace - Crisp support-thread intake - Read-only GitHub repo context - Sentry runtime evidence Business For steady support load - Everything in Starter - Up to 150 investigations - Priority support - Faster tuning Enterprise For larger rollouts - Higher-volume queues - Multiple workspaces - Priority rollout help - Custom terms Free trial includes 5 investigations • monthly only • priced per workspace Frequently Asked Questions How does the free trial work? Your first 5 investigations are included. Start from signup, then connect your support workspace, repo context, and runtime signals. Can I trust the output? Rinhelp produces review drafts only when the evidence is strong enough. When context is missing, it shows the gap and the exact information to collect next. What about security? The current product boundary is narrow on purpose: tenant-scoped Crisp, GitHub, and Sentry connections, read-only repo and runtime access, and no customer-facing auto replies. Why not just use Codex, Claude Code, or Cursor? Those tools can help once you steer them. Rinhelp starts from the support thread, gathers the technical context, and runs the first investigation pass for you. Keep building until it is time to review the diagnosis If technical support diagnosis keeps stealing the time and attention you meant to spend shipping, Rinhelp gives you an evidence-backed investigation workflow and a human review loop before anything goes out. START FREE TRIALFor founder-led and engineer-led SaaS teams with real technical ticket volume
Genesis Park 편집팀이 AI를 활용하여 작성한 분석입니다. 원문은 출처 링크를 통해 확인할 수 있습니다.
공유