Instagram Reels API audio: AAC, 48 kHz max, mono or stereo

For an Instagram Reel through the API, Meta's IG User Media reference lists the audio as AAC, 48 kHz sample rate maximum, 1 or 2 channels (mono or stereo), with an audio bitrate of 128 kbps (read 2026-10-02). Sume's video inspect probe returns audio_codec, audio_channels, audio_sample_rate and has_audio, which covers three of those four items before you create a container.

What does the Reel audio spec say?

The audio line sits next to the video requirements in the Reel specifications. Stories use the same audio line.

Which of these can I check with Sume?

Codec, channel count and sample rate. The probe schema has no audio bitrate field, so that item stays unchecked. Probe and stills are unbilled; with frames: false you get the probe alone.

A clip without sound reports has_audio: false and a null codec. The page does not say whether a Reel must have audio, so a silent clip is something to test, not to assume.

import hashlib, os, time, requests
BASE = "https://api.sume.com"
H = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"}

def probe(video_url):
    key = "probe-" + hashlib.sha256(video_url.encode()).hexdigest()[:24]
    r = requests.post(f"{BASE}/v1/video-inspect", headers={**H, "Idempotency-Key": key},
                      json={"video_url": video_url, "frames": False}, timeout=60)
    r.raise_for_status()
    rid = r.json()["data"]["video_inspect_id"]
    while True:
        d = requests.get(f"{BASE}/v1/video-inspect/{rid}", headers=H, timeout=30).json()["data"]
        if d["resource_status"] == "ready":
            return d["probe"]
        if d["resource_status"] in ("failed", "canceled"):
            raise RuntimeError(d["error"])
        time.sleep(2)

p = probe("https://media.sume.com/artifacts/artf_demo/talk.mp4")
problems = []
if not p["has_audio"]:
    print("no audio track")
else:
    if (p["audio_codec"] or "").lower() != "aac": problems.append(f"codec {p['audio_codec']}")
    if (p["audio_sample_rate"] or 0) > 48000: problems.append(f"{p['audio_sample_rate']} Hz")
    if p["audio_channels"] not in (1, 2): problems.append(f"{p['audio_channels']} channels")
print(problems or "audio fields within the listed limits")

What if the audio is outside the limits?

Video trim with precision: exact re-encodes the clip and, per its docs, remuxes kept audio as AAC. That fixes a codec mismatch. The docs do not list a sample-rate or channel option on trim, so a 96 kHz or 6-channel source needs another route; regenerate the audio at a supported shape instead.

Trim's precision: keyframe is a stream copy, so it keeps the source audio unchanged.

What should I do?

Probe every file, fail on the three checkable fields, and keep the 128 kbps item as a manual check on a sample. The Facebook Reels audio spec is covered in its own post; this one is Instagram only.