
LastSearch
Research with evidence.
Real-time web search for agents where every claim carries a source and every answer carries a confidence score that was computed, not self-reported.
npx lastsearch
Every tool shipped so far. Open source, free, and built on one rule: nothing is claimed without proof.

Research with evidence.
Real-time web search for agents where every claim carries a source and every answer carries a confidence score that was computed, not self-reported.
npx lastsearch

Runtime observation.
Attach to any running process, container or pod and turn strace, py-spy, perf, lsof and heap snapshots into findings that cite evidence ids.
npx debuganything why <pid>

Watch the work.
A coding-agent session rendered as a film instead of a log. Every read, edit and test run becomes a visible beat, and a claim of success with no passing test behind it is flagged on screen.
npx agent-theatre

A working twin of any API.
A local, stateful copy of any API on your machine: objects persist, statuses move through real lifecycles, webhooks are signed the way the provider signs them, and you can force declines, timeouts and outages on demand. For your app and your agents; nothing leaves your machine.
npx @shipatmidnight/mockanything stripe

Is this true?
Give it a claim, a paragraph, a URL, a file or an AI answer. It pulls out every factual claim and judges each against live sources: verified, contradicted or unverified, with a confidence estimate, the exact evidence and the reason. A CLI and an MCP server, powered by LastSearch.
npx verify-anything "The Eiffel Tower is in Berlin."

When CI hangs, get stacks instead of a timeout.
A GitHub Action that wraps your test command. If it finishes, nothing changes. If it hangs, Debug Anything attaches before the job is killed and uploads every thread's stack, socket and file states, and a ranked report where every finding cites its evidence.
uses: shreyassaw5394/debuganything-action@v1