Reka Core — Reka ingests video, audio, and image in one call. peepshow caps cost on long footage and slots in animated formats.

Reka already reads video. Why peepshow?

Reka accepts video natively. peepshow is not a competing video understanding stack — it's a control plane that sits in front of native video to handle the cases native upload doesn't.

• Cost ceiling on long clips. Reka's native video bills per-second. peepshow extracts N scene-change frames so a 1-hour CCTV reel costs predictably instead of scaling linearly with seconds.
• Animated GIF / APNG / WebP. Reka's video path expects container formats. peepshow flattens animated images into a clean frame sequence.
• Audit trail. Reka's native sampling is opaque. peepshow gives you the exact frames the model saw — replayable, diffable, cacheable.
• Audio transcript stays cheap. whisper.cpp on PATH → plain text transcript. Cheaper than feeding the raw audio for dialogue-heavy content.
• Cross-model portability. Same peepshow bundle feeds Reka + Gemini + Claude + GPT. Extract once, run anywhere.
• Sinks. Reka won't push frames to Notion / Slack / S3 / SQL. peepshow does — same run, 95 destinations.

Token-cost math (worked examples)

Reka publishes per-second video pricing across Core / Flash / Edge. Numbers above are approximate against the current Core tier — Flash and Edge are cheaper but with smaller context.

Install (CLI)

npm install -g peepshow

# Set Reka credentials:
export REKA_API_KEY=...

# Extract — peepshow's manifest still works alongside Reka's native video:
peepshow ./demo.mp4 --emit json > run.json

Install (Reka API directly, no CLI)

Calling the Reka API from your own code? Run peepshow first, then feed the JSON manifest in as multimodal parts:

# Hand frames + transcript to Reka (OpenAI-compatible chat endpoint)
node -e '
  import OpenAI from "openai";
  import { readFileSync } from "node:fs";
  const run = JSON.parse(readFileSync("run.json", "utf8"));
  const content = [
    { type: "text", text: "Summarise this clip." },
    ...run.frames.map(f => ({
      type: "image_url",
      image_url: { url: "data:image/jpeg;base64," + readFileSync(f.path).toString("base64") }
    })),
    { type: "text", text: "Transcript:\n" + (run.transcript?.text ?? "") }
  ];
  const client = new OpenAI({
    apiKey: process.env.REKA_API_KEY,
    baseURL: "https://api.reka.ai/v1",
  });
  const r = await client.chat.completions.create({
    model: "reka-core",
    messages: [{ role: "user", content }]
  });
  console.log(r.choices[0].message.content);
'

Animated GIF / APNG / WebP — peepshow's killer move on Reka

Reka's native video path expects video containers. Animated PNG, animated WebP, and most GIFs come back as still images — motion lost. peepshow probes the source, treats it as a frame sequence, and emits ordinary JPEGs Reka will gladly read alongside its native modalities.

peepshow ./meme.gif         # animated GIF → frame timeline
peepshow ./tutorial.apng    # animated PNG → frames
peepshow ./loop.webp        # animated WebP → frames

Frame strategy presets

peepshow picks scene-change frames by default. For Reka specifically, these presets are worth knowing:

• --strategy scene --max 20Default — good balance for narrative clips that exceed Reka's native short-clip sweet spot.
• --strategy fps --fps 1 --max 60Match Reka's internal video sampler cadence for comparison runs.
• --strategy scene --max 12 --dedup perceptualLong static footage (CCTV / timelapse). Drops near-duplicates before Reka pays for them.

All 95 sinks still fire

Same CLI = same sinks. Push frames to SQLite, embed captions into Chroma, mirror to S3, drop a thumbnail in Slack, file a GitHub issue with the offending frame attached — all from one Reka run. Browse the full sink catalogue →.

Report + LLM analysis loop

Every run also writes a self-contained report.html + manifest.json next to the frames (see the Report page). When Reka consumes the frames, the analysis flows back into the report — whoever opens it next sees the model's understanding without re-running the prompt.

echo '{"summary":"<Reka's summary>","provider":"reka-core"}' \
  | peepshow report annotate "<outputDir>"

When to skip peepshow + use Reka direct

• Clip is under 60s and you want full native motion + audio fidelity (lip-sync, sport, music).
• One-shot prototype — token cost doesn't matter, and you don't need the audit trail.
• No need to fan out to other models or sinks — Reka's native path is fine on its own.

Native video is excellent for those cases. peepshow earns its place when any of: long, animated format, cost-bounded, auditable, multi-model, local-first, fan out to other systems.