Published 2026-09-22 · Tested 2026-09-22

Mux

C

Mux received 6 PASS votes and passed 4 of five agent surface checks. The clearest finding came from the find the exact limits task.

Panel: GPT 5.6 Sol, Opus 5, DeepSeek v4F Battery: v1 Read as markdown (opens in a new tab)

Three AI models, GPT 5.6 Sol, Claude Opus 5, and DeepSeek v4 Flash, each read Mux’s public documentation independently and attempted five first-hour developer jobs: create the first asset, find the exact limits, recover from a 429, verify a webhook, use the Node SDK.

No accounts, API calls, or code execution were used. Every verdict came from public pages and every published quotation passed a live verification check.

Freshness

How rechecks work
Tested
Quotes verified
Surface rechecked
Not yet rechecked
Mux Custom Next.js · published
C

76.2% · 61/80 · AI Agent Readiness Score · reading 30 pts · surface 50 pts

llms.txt PASS
llms-full.txt PASS
markdown mirror PASS
MCP server PASS
docs AI FAIL
Task GPT 5.6 SolOpus 5DeepSeek v4F Consensus
Create the first asset PASSPASSPASS PASS
Find the exact limits PARTIALPARTIALPARTIAL PARTIAL
Recover from a 429 PARTIALPARTIALPARTIAL PARTIAL
Verify a webhook PASSPASSPASS PASS
Use the Node SDK PARTIALPARTIALPARTIAL PARTIAL

docs platform: Custom Next.js (unscored) · verified 2026-09-22

What to fix first

These 4 fixes could add up to 19 points to the AI Agent Readiness Score. The list ranks each fix by the points it would add. How the ranking works

  1. 1
    +10 points Docs AI check failed

    Found: The hydrated docs home offers only a keyboard search control, and no assistant control or assistant provider script loads.

    Fix: Add an assistant to the docs site that answers questions from the docs and links to its sources.

  2. 2
    +3 points Find the exact limits PARTIAL

    Found: The API page states the request-bucket limits but leaves the organization-wide cap unquantified, and plan quotas such as 10 stored videos appear only on the pricing page.

    Fix: On the make-API-requests page, state the organization-wide cap as a number and link the plan quotas, including the Free plan's 10 stored videos.

    Evidence: www.mux.com/docs/core/make-api-requests (opens in a new tab)

  3. 3
    +3 points Recover from a 429 PARTIAL

    Found: The docs name only x-ratelimit-limit and x-ratelimit-remaining and document no reset or Retry-After header, so an agent must compute the wait after a 429 by hand.

    Fix: On the make-API-requests page, add a worked 429 retry example that picks the bucket from the request method and computes the wait from the fill-rate table.

    Evidence: www.mux.com/docs/core/make-api-requests (opens in a new tab)

  4. 4
    +3 points Use the Node SDK PARTIAL

    Found: The Node SDK example creates an asset with input and playback_policy, fields the API reference schema marks deprecated and replaces with inputs and playback_policies.

    Fix: Rewrite the Node SDK create-asset example with inputs and playback_policies so it matches the create-asset reference schema.

    Evidence: www.mux.com/docs/integrations/mux-node-sdk (opens in a new tab)

What the docs get right

  • Create the first asset: 3 PASS votes. One quickstart covers token creation, URL ingest, readiness, the playback ID, and browser playback.
  • Verify a webhook: 3 PASS votes. Creation, delivery, and HMAC-SHA256 verification are complete; the webhook-reference fetch exceeded the tool's 5 MB limit.
  • 4 of 5 agent surface checks. Present: llms.txt, llms-full.txt, markdown mirrors, an MCP server. Missing: docs AI.

Create the first asset

PASS

PASS consensus from 3 PASS.

One quickstart covers token creation, URL ingest, readiness, the playback ID, and browser playback. One numbered page carries token generation with the required Mux Video Read and Write permissions, a runnable curl POST whose rendered body is {"inputs": [{"url": ...}]}, demo source URLs, the preparing-to-ready status wait, and the https://stream.mux.com/{PLAYBACK_ID}.m3u8 playback URL. The guide steps from Access Token creation through POST /assets to playback URL construction in one linear path; a minor discrepancy (guide says singular input but REST API uses plural inputs) does not block the path.

Find the exact limits

PARTIAL

PARTIAL consensus from 3 PARTIAL.

API and pricing values agree, but the organization-wide cap and Data burst sizes remain unquantified. The API rate limits are stated exactly as a four-row table (POST high 20 requests at 1 req/sec, POST low 4 at 1 req/sec, all other methods high 100 at 5 req/sec, low 20 at 1 req/sec) plus 1 RPS sustained for Monitoring Data and 5 RPS sustained for General Data, but plan quotas such as the Free plan's "Up to 10 videos stored" appear only on the pricing page, and a search of the whole llms-full.txt docs corpus returns zero matches for quota, usage limit, or free trial. API request-bucket rate limits are documented with exact numbers; account-level plan quotas (10 stored videos on Free, 100K delivery minutes) are only on the separate pricing page with no cross-reference from the API docs, forcing an agent to find and reconcile two documentation areas.

Recover from a 429

PARTIAL

PARTIAL consensus from 3 PARTIAL.

The headers expose capacity and remaining requests, but no reset or retry header gives an exact retry time. The docs name only x-ratelimit-limit and x-ratelimit-remaining, document no Retry-After or reset header, and give no backoff algorithm, so the wait must be computed by hand from the fill-rate table after inferring the bucket from the request method. The x-ratelimit-limit and x-ratelimit-remaining headers are documented and the fill-rate table gives the refill speed, but no Retry-After header is mentioned and no explicit retry algorithm (e.g. how long to wait, exponential backoff) is provided; agents must infer the delay from the bucket size and fill rate.

Verify a webhook

PASS

PASS consensus from 3 PASS.

Creation, delivery, and HMAC-SHA256 verification are complete; the webhook-reference fetch exceeded the tool's 5 MB limit. The path is complete end to end: create the endpoint in the Dashboard or with a documented curl POST to https://api.mux.com/system/v1/webhooks that returns a one-time signing_secret, receive the POST within the stated 5 second timeout, tunnel locally with the Mux CLI, then follow four numbered verification steps and the 5 minute default timestamp tolerance. Full end-to-end path: create webhooks via Dashboard or API, signing secret delivered in create response, mux-signature header format documented, HMAC-SHA256 verification steps shown, Node SDK provides mux.webhooks.unwrap helper; webhook events guide links directly to verification guide.

Use the Node SDK

PARTIAL

PARTIAL consensus from 3 PARTIAL.

The npm fetch returned 403, the API Markdown fetch returned 404, and the integration page conflicts with current sources. The SDK is easy to find and the example runs, but it uses input and playback_policy, which the API reference schema marks deprecated in favour of the inputs and playback_policies fields that the same site's quickstart curl and the create-asset request example both use. The Node SDK (@mux/mux-node) has a working example on GitHub that uses inputs (plural) matching the REST API OpenAPI spec example; however the Mux docs integration page for Node returns HTTP 404, and the quickstart guide describes setting input (singular) while the SDK and REST API both require inputs (plural array), creating a cross-reference gap.

The receipt

Each environment also counts toward an organization-wide cap.

API and pricing values agree, but the organization-wide cap and Data burst sizes remain unquantified.

Agent surface notes

Initialize returned HTTP 401 with a Bearer OAuth challenge naming resource metadata at mcp.mux.com/.well-known/oauth-protected-resource.

The hydrated docs home offers only a keyboard search control, and no assistant control or assistant provider script loads.

Show the score

AI Agent Readiness Score 76.2%, grade C

Paste this into a readme:

[![AI Agent Readiness Score 76.2%](https://docsforagents.com/badge/mux.svg)](https://docsforagents.com/reports/mux-docs-ai-agent-readiness/)

Method note

This is a reading test of public documentation, not an execution test. No accounts were created and no API calls were run. The AI Agent Readiness Score counts fifteen reading votes at PASS 2, PARTIAL 1, and FAIL 0, for 30 possible points. Five agent surface checks add 10 points each. The total is 80. Consensus chips show each row majority and do not affect scoring. The panel split on 0 of five tasks. Quotes shown here were re-fetched and confirmed verbatim on 2026-09-22.

Read the full methodology

Put another docs site through the battery.

Nominate a docs site