ReelQL
Claude skill and API that turns a video link into timestamped JSON, then asks Jev typed questions to rank or filter videos against a brief.
Overview
ReelQL watches a video from a link and returns one typed JSON document: summary, chapters, scenes, cast, key moments, emotional arc, products and brands, on-screen text and the full transcript, all timestamped. The README reports link-to-JSON times from 16 s for a 6-second TikTok to 65 s for a 14-minute walking tour.
Jev reads text only, so ReelQL is the step that turns video into fields Jev can judge. The bundled Claude skill runs each link through ReelQL and then a script asks Jev five questions per video: on brief, hook, product in use, emotional payoff and unsafe content. Your code then sorts, filters or flags on the probabilities; in the README's example, five brand TikToks are ranked against a family-friendly brief. Recipes cover brand safety and choosing the best moment for a cut-down. The README advises sending Jev only the fields a question needs because long transcripts can exceed its context, and doing numeric comparisons in code. A ReelQL key comes with 10 free minutes; the Jev key is optional.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Does `transcript` contain profanity, violence or adult content? | — |
| Yes / no | Does the video match `brief`? | — |
| Score | How well does `tone` fit `brief`? | Contradicts it · Neutral · Clearly fits |
| Choice | Which key moment best matches the brief? | — |
How it works
- Send a video link to ReelQL
- Receive a timestamped JSON analysis
- Pass only the needed fields to Jev with typed questions
- Rank, filter or route videos in code on the returned probabilities
Strengths
- Makes video judgeable by a text-only decision model
- Probabilities rather than prose to parse
- Example outputs available without a key
Caveats
- Video analysis runs on ReelQL's hosted service
- Long transcripts can exceed Jev's context
- Not affiliated with TypeSafe
Getting started
Get a key with curl -s -X POST https://reelql.tail6c0e2d.ts.net/keys, install the skill, and optionally export TYPESAFE_API_KEY for Jev rankings.At a glance
- Type
- Agent skill
- Status
- beta
- Language
- Python
- Runtime
- Claude skill + hosted API
- Has code
- yes
Use cases and integrations
How curated lists describe it
Claude skill and API that turns a video link into one typed JSON document: chapters, scenes, key moments, on-screen text and transcript.
MrJev/awesome-jev
Frequently asked questions
What does Jev decide in ReelQL?
Whether a video is on brief and brand-safe, how well its tone fits, and which moment best matches a brief.
Why not send the video to Jev directly?
Jev reads text only, so ReelQL first converts the video into structured fields.
Is a TypeSafe key required?
Only for the Jev rankings and judgments; ReelQL analysis needs its own key.
Related projects
Video summarizer that ties every line to its second; Jev checks each summary sentence against what was said, with numbers checked in code.
ProjectbetaPythonLocal long-term memory for coding agents with an optional Jev reranker that orders recalled memories.
ProjectproductionTypeScriptAgent skill and Python library for Claude Code and Codex that reranks candidate lists with one Jev yes/no relevance question each.
ProjectbetaPythonMCP server that scores every search query, result and fetched page an agent sees for relevance, credibility and steering risk.
MCP serverbetaTypeScriptCLI that scores every test against a git diff with Jev and emits filter arguments for vitest, jest, node, Playwright, cargo, go and rspec.
ProjectbetaTypeScriptSingle MCP tool that indexes a repo and has Jev rank a shortlist so a coding agent reads the right files and lines instead of grepping.
MCP serverbetaPython