Install Jentic One Beta
Jentic One is a self-hosted execution layer for AI agents. It lets your agent call the SafeReel, or any other public or private API you need. You set the rules, the agent never sees your credentials, and every call is logged.
Two steps, two machines. Install the instance in a safe environment, then register your agent from wherever it runs.
Step 1: Jentic One Host machine
# On the machine that will host your Jentic One instance:
curl -fsSL "https://jentic.com/install.sh?src=apis&api=%2Fapis%2Fsafereel.ai%2Fsafereel" | shStep 2: Agent machine
# On the machine where your agent runs (keep this separate from the instance):
curl -fsSL "https://jentic.com/install.sh?src=apis&api=%2Fapis%2Fsafereel.ai%2Fsafereel" | sh
jentic register # connects your agent to your Jentic One instanceJentic One is in public beta. The setup above keeps your agent separate from the instance, which is what you want before using real credentials: an agent running as the same OS user as Jentic One can read its stored keys directly. Just evaluating? A single local install is fine to start. See the secure deployment guide for the tiers.
What an agent can do with SafeReel API.
Classify one uploaded JPEG, PNG, or WebP image as nsfw or clean with a confidence score
Queue a video scan from a public URL and receive a job id to poll
Receive a signed webhook callback when a video scan finishes by setting webhook_url
Read a video job's status, verdict, flagged frame count, and flagged_ts timestamps
Make video submissions safe to retry with an Idempotency-Key header
GET STARTED
Check public service status and load before sending a large batch
Patterns agents use SafeReel API for, with concrete tasks.
★ Upload moderation agent through Jentic
A community app wants an agent to screen new posts before they publish. Through Jentic the agent searches for 'check an image for NSFW content', receives the SafeReel operation with its input schema, and calls POST /v1/check with the uploaded image. It reads the verdict and confidence and holds any nsfw post for review, while clean posts go live.
Call POST /v1/check with the uploaded image file, then publish the post if verdict is 'clean' and hold it for review if verdict is 'nsfw' with confidence above 0.8
Asynchronous video screening
Video platforms cannot block an upload request while a long clip is scanned. POST /v1/videos takes a public url and an optional webhook_url and returns a job id with status. GET /v1/videos/{job_id} reports queued, processing, done, or failed, and once done gives the verdict, the number of frames and flagged frames, and flagged_ts with the timestamps of flagged frames.
Call POST /v1/videos with url 'https://cdn.example.com/uploads/clip.mp4' and an Idempotency-Key header, then poll GET /v1/videos/{job_id} until status is 'done' and report flagged_ts
Webhook-driven moderation pipeline
Instead of polling, a backend can pass webhook_url when submitting a video. The submit response includes a webhook_secret, returned only at submit time, that signs the webhook HMAC so the receiver can verify the callback came from SafeReel before acting on the verdict.
Call POST /v1/videos with url 'https://cdn.example.com/uploads/clip.mp4' and webhook_url 'https://example.com/hooks/safereel', and store the returned webhook_secret for HMAC verification
6 endpoints — the safereel api classifies images and videos as nsfw or clean so platforms can screen user uploads before they go live.
METHOD
PATH
DESCRIPTION
/v1/check
Classify one image
/v1/videos
Submit a video for scanning
/v1/videos/{job_id}
Poll a video job
/v1/status
Public service status
/v1/demo/check
Anonymous demo image check
/v1/check
Classify one image
/v1/videos
Submit a video for scanning
/v1/videos/{job_id}
Poll a video job
/v1/status
Public service status
/v1/demo/check
Anonymous demo image check
What agents get from Jentic-routed access to this vendor.
Setup
Wiring SafeReel by hand means sending a bearer key, handling a synchronous image path and an asynchronous video path, polling jobs or verifying signed webhooks, and adding an Idempotency-Key for safe retries. Through Jentic you install once, import SafeReel from the API Directory, store the key once, and your agent calls it.
Permission scoping
SafeReel puts the video job id in the URL path (/v1/videos/{job_id}), so a rule can let your agent read one job's result, while image checks are bounded by operation. You choose the operations it may call, so billable video submissions are not included unless you add them.
Credential isolation
Your SafeReel API key is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context.
Intent-based discovery
Agents search Jentic by intent such as 'check an image for NSFW content' or 'scan a video for explicit content', and Jentic returns the matching SafeReel operation with its input schema so the agent calls the right endpoint without reading the docs.
Alternatives and complements available in the Jentic catalogue.
Specific to using SafeReel API through Jentic.
What authentication does the SafeReel API use?
The SafeReel API declares an HTTP bearer scheme named HTTPBearer in its OpenAPI spec, used by POST /v1/check, POST /v1/videos, and GET /v1/videos/{job_id}. GET /v1/status, GET /health, and the POST /v1/demo/check demo need no auth. Through Jentic the bearer key is held by your own Jentic One instance and attached at call time.
Can I scan long videos with the SafeReel API?
Yes. POST /v1/videos queues a scan from a public http or https URL with no upload and no file-size limit, and the spec notes every video bills at most 20 minutes. Poll GET /v1/videos/{job_id} or set webhook_url for a signed callback. Job ids are owner-only, so other keys get a 404.
What are the rate limits for the SafeReel API?
The spec says POST /v1/check and the demo report rate-limit state in RateLimit-* response headers, and the anonymous demo is limited to 10 images per IP per day. It gives no fixed limit for keyed calls; check https://safereel.ai/docs for current plan limits.
How do I check an image for NSFW content through Jentic?
Search Jentic for 'check an image for NSFW content', load POST /v1/check, and execute it with a JPEG, PNG, or WebP image of up to 10MB. The response gives verdict, confidence, model, and whether the result was cached. To run it on your own infrastructure, install Jentic One from its GitHub repo.
Is there a SafeReel MCP server?
You don't need an MCP server to give your agent SafeReel. Jentic connects it directly from the API Directory: import it, store your API key once, and your agent finds the image check and video scan operations when it needs them.
Can I limit what my agent is allowed to do with the SafeReel API?
Yes. For an image-only moderator, write a rule that allows only POST /v1/check. The agent can classify images but cannot queue billable video scans, and each call it makes is logged. Add POST /v1/videos and GET /v1/videos/{job_id} when you want it to handle video.
For Agents
Check an image for NSFW content and get a verdict with confidence, or queue a video scan by URL and read flagged frames and timestamps when the job is done.
Use for: Check whether this uploaded image is NSFW, I need to scan a user-submitted video for explicit content, Get the result of my SafeReel video scan job, Find the timestamps where a video was flagged as NSFW
Not supported: Does not detect violence, hate speech, or spam text, and does not host or transcode media. Use for nsfw or clean classification of images and videos only.
The SafeReel API classifies images and videos as nsfw or clean so platforms can screen user uploads before they go live. Images get a synchronous verdict with a confidence score from POST /v1/check, and cached verdicts return instantly without being billed. Videos are scanned asynchronously from a public URL with no upload or file-size limit, returning a job to poll or a signed webhook callback, and the finished job reports how many frames were flagged and at which timestamps.
This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use.
Base layer of spec validity and structural soundness.
Aggregated quality score from linter diagnostics, weighted by severity.
Percentage of `$ref` references that resolve successfully.
Checks whether the API description parses successfully and conforms to its declared specification (e.g., OpenAPI).
Structural correctness score based on schema issues using logarithmic dampening.
Clarity, completeness, and ingestion readiness for developers and tooling.
How richly the API is illustrated with examples.
Percentage of examples that conform to their schemas.
Percentage of operations with complete response definitions (success, client error, server error).
Health of API ingestion, bundling, and resolution within Jentic pipelines.
Semantic breadth, depth, and agent comprehension for AI systems.
Coverage of descriptions across API elements.
Coverage of RFC 9457 Problem Details for error responses.
Coverage, uniqueness, and casing consistency of operationIds for AI inference.
Coverage of summaries across operations/tags/info.
Functional utility, complexity comfort, and AI orchestration readiness.
Agent comfort level based on API operational and structural complexity.
Trust, risk posture, and security compliance.
Average quality of security schemes based on authentication method strength (weakest link for OAuth2).
Findability, semantic richness, and reasoning readiness.
Clarity and depth of descriptions across API elements.
Score it yourself
Every API in the directory is allowlisted, so you can re-score it with no key required.
npx @jentic/api-scorecard-cli score <openapi-url>