Resemble Detect

Resemble Detect — Deepfake Detection & Media Safety

Analyze audio, image, video, and text for synthetic manipulation, AI-generated content, watermarks, speaker identity, and media intelligence using the Resemble AI platform.

Core Principle — THE IRON LAW

"NEVER DECLARE MEDIA AS REAL OR FAKE WITHOUT A COMPLETED DETECTION RESULT."

Do not guess, infer, or speculate about media authenticity. Every authenticity claim must be backed by a completed Resemble detect job with a returned label, score, and status: "completed". If the detection is still processing, wait. If it failed, say so — do not substitute your own judgment.

When to Use

Use this skill whenever the user's request involves any of these:

• Checking if audio, video, image, or text is AI-generated or manipulated
• Detecting deepfakes in any media format
• Verifying media authenticity or provenance
• Identifying which AI platform synthesized audio (source tracing)
• Applying or detecting watermarks on media
• Analyzing media for speaker info, emotion, transcription, or misinformation
• Asking natural-language questions about detection results
• Matching or verifying speaker identity against known voice profiles
• Detecting AI-generated or machine-written text
• Any mention of: "deepfake", "fake detection", "synthetic media", "voice verification", "watermark", "media forensics", "authenticity check", "source tracing", "is this real", "AI-written text", "text detection"

Do NOT use for text-to-speech generation, voice cloning, or speech-to-text transcription — those are separate Resemble capabilities.

Capability Decision Tree

When multiple capabilities apply (e.g., user wants deepfake detection AND intelligence), combine them in a single POST /detect call using the intelligence: true flag rather than making separate requests.

Required Setup

• API Key: Bearer token from the Resemble AI dashboard (set as RESEMBLE_API_KEY)
• Base URL: https://app.resemble.ai/api/v2
• Auth Header: Authorization: Bearer <RESEMBLE_API_KEY>
• Media Requirement: All media must be at a publicly accessible HTTPS URL

If the user provides a local file path instead of a URL, inform them the file must be hosted at a public HTTPS URL first. Do not attempt to upload local files to the API. (Exception: POST /text_detect accepts text content inline.)

MCP Tools Available

When the Resemble MCP server is connected, use these tools instead of raw API calls:

Tool usage pattern: Use resemble_docs_lookup with topic "detect" to get the full picture, then resemble_api_endpoint for exact request/response schemas before making API calls.

Full API Reference

Detailed request/response schemas for every endpoint are in references/api-reference.md. Consult it before making any API call to verify exact parameter names and response shapes. The sections below cover decision-making; the reference covers exact field formats.

---

Phase 1: Deepfake Detection

The core capability. Submit audio, image, or video for AI-generated content analysis via POST /detect.

Key flags to consider:

• visualize: true — generate heatmap/visualization artifacts
• intelligence: true — run multimodal intelligence alongside detection (saves a round-trip)
• audio_source_tracing: true — identify which AI platform synthesized fake audio (only fires on "fake" audio)
• use_reverse_search: true — enable reverse image search (image only)
• zero_retention_mode: true — auto-delete media after analysis (for sensitive content)

Detection is asynchronous. Poll GET /detect/{uuid} at 2s → 5s → 10s intervals until status is "completed" or "failed". Most complete in 10–60 seconds.

Supported formats: Audio (WAV, MP3, OGG, M4A, FLAC) · Video (MP4, MOV, AVI, WMV) · Image (JPG, PNG, GIF, WEBP)

Reading Results

• Audio — verdict in metrics — use label and aggregated_score
• Image — verdict in image_metrics — use label and score; ifl has an Invisible Frequency Layer heatmap
• Video — verdict in video_metrics — hierarchical tree of frame/segment results; video-with-audio returns both metrics and video_metrics

See references/api-reference.md for full response schemas.

Interpreting Scores

Always present scores with context. Say "The detection returned a score of 0.87, indicating high confidence that this audio is AI-generated" — never just "it's fake."

---

Phase 2: Intelligence — Media Analysis

Rich structured insights about media: speaker info, emotion, transcription, translation, misinformation, abnormalities.

Two ways to run Intelligence:

1. Combined with detection — add intelligence: true to POST /detect (preferred; one call)
2. Standalone — POST /intelligence with a URL (when you only need analysis, not a deepfake verdict)

Audio/video structured fields include: speaker_info, language, dialect, emotion, speaking_style, context, message, abnormalities, transcription, translation, misinformation.

Image structured fields include: scene_description, subjects, authenticity_analysis, context_and_setting, abnormalities, misinformation.

Detect Intelligence — Ask Questions About Results

After a detection completes, ask natural-language questions via POST /detects/{detect_uuid}/intelligence with { "query": "..." }. Returns a question UUID — poll GET /detects/{detect_uuid}/intelligence/{question_uuid} until completed.

Good questions to suggest:

• "Summarize the detection results in plain language"
• "What specific indicators suggest this is AI-generated?"
• "How do the audio and video detection results differ?"
• "What is the confidence level and what does it mean?"
• "Are there any inconsistencies in the analysis?"

Prerequisite: The detection must have status: "completed". Submitting a question against a processing or failed detection returns 422.

See references/api-reference.md for full parameters.

---

Phase 3: Audio Source Tracing

When audio is labeled "fake", identify which AI platform generated it.

Enable it by setting audio_source_tracing: true in the POST /detect request. Result appears in the detection response under audio_source_tracing.label.

Known labels: resemble_ai, elevenlabs, real, and others as the model expands.

Important: Source tracing only runs on audio labeled "fake". Real audio produces no source tracing result.

Standalone queries: GET /audio_source_tracings and GET /audio_source_tracings/{uuid}.

---

Phase 4: Watermarking

Apply invisible watermarks to media for provenance tracking, or detect existing watermarks.

• Apply: POST /watermark/apply with url, optional strength (0.0–1.0), optional custom_message. Add Prefer: wait for synchronous response, or poll GET /watermark/apply/{uuid}/result. Response includes watermarked_media URL.
• Detect: POST /watermark/detect with url. Audio returns { has_watermark, confidence }; image/video returns { has_watermark }.

See references/api-reference.md for exact parameter rules.

---