# Honeycomb docs: AI Agent Readiness Score 87.5% (B+)

**87.5% · 70/80 · AI Agent Readiness Score · 20/30 reading points · 50/50 agent surface points**

Honeycomb received 6 PASS votes and passed 5 of five agent surface checks. The clearest finding came from the find the exact limits task.

- Tested: 2026-09-27
- Published: 2026-09-27
- Battery: v1
- Scoring: reading 30 pts · surface 50 pts
- Docs: https://docs.honeycomb.io/

Three AI models, GPT 5.6 Sol, Claude Opus 5, and DeepSeek v4 Flash, each read Honeycomb’s public documentation independently and attempted five first-hour developer jobs: send the first telemetry, find the exact limits, recover from a 429, verify a webhook, use the Web 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](https://docsforagents.com/methodology/#freshness)

- Category: [Observability & analytics](https://docsforagents.com/grades/?category=observability)
- Tested: 2026-09-27
- Quotes verified: 2026-09-27
- Surface rechecked: Not yet rechecked

## Agent surface checks · 50/50

| Check | Verdict | Points |
| --- | --- | --- |
| llms.txt | PASS | 10 |
| llms-full.txt | PASS | 10 |
| Markdown mirror | PASS | 10 |
| MCP server | PASS | 10 |
| Docs AI | PASS | 10 |

## The Reading Test · 20/30

| Task | GPT 5.6 Sol | Opus 5 | DeepSeek v4F | Consensus |
| --- | --- | --- | --- | --- |
| Send the first telemetry | PASS | PASS | PASS | PASS |
| Find the exact limits | FAIL | PARTIAL | PARTIAL | PARTIAL |
| Recover from a 429 | PARTIAL | PARTIAL | PASS | PARTIAL |
| Verify a webhook | PASS | PARTIAL | PASS | PASS |
| Use the Web SDK | PARTIAL | PARTIAL | PARTIAL | PARTIAL |

Docs platform: Mintlify (unscored) · verified 2026-09-27

## What to fix first

These 3 fixes could add up to 9 points to the AI Agent Readiness Score. The list ranks each fix by the points it would add. [How the ranking works](https://docsforagents.com/methodology/#what-to-fix-first)

1. **+4 points · Find the exact limits · PARTIAL**

   **Found:** The Rate Limits page describes the general API rate limit but never states its numeric value, window, or scope.

   **Fix:** State the general API rate limit's value, window, and scope on the Rate Limits page.

   **Evidence:** [docs.honeycomb.io/api/rate-limit](https://docs.honeycomb.io/api/rate-limit)

2. **+3 points · Use the Web SDK · PARTIAL**

   **Found:** The Web SDK guide installs @honeycombio/opentelemetry-web, but its verification step names @honeycomb/opentelemetry-web.

   **Fix:** Correct the verification step to @honeycombio/opentelemetry-web and use one EU endpoint value on both Web SDK pages.

   **Evidence:** [docs.honeycomb.io/send-data/javascript-browser](https://docs.honeycomb.io/send-data/javascript-browser)

3. **+2 points · Recover from a 429 · PARTIAL**

   **Found:** The Rate Limits page promises Problem Detail bodies for V1, but the Create an Event reference shows a bare object.

   **Fix:** Reconcile the Create an Event 429 body example with the V1 Problem Detail format on the Rate Limits page.

   **Evidence:** [docs.honeycomb.io/api/rate-limit](https://docs.honeycomb.io/api/rate-limit)

## What the docs get right

- **Send the first telemetry: 3 PASS votes.** The quickstart links signup, API-key access, OpenTelemetry setup, requests, and Honeycomb UI confirmation in one sequence.
- **Verify a webhook: 2 PASS votes.** One guide covers recipient creation, Trigger attachment, delivery, and matching X-Honeycomb-Webhook-Token against the Shared Secret.
- **5 of 5 agent surface checks.** Present: llms.txt, llms-full.txt, markdown mirrors, an MCP server, docs AI.

## Send the first telemetry

**PASS**

PASS consensus from 3 PASS.

The quickstart links signup, API-key access, OpenTelemetry setup, requests, and Honeycomb UI confirmation in one sequence. One page chains signup region choice, team creation, API key capture, per-language automatic instrumentation, running the app, and a HEATMAP(duration_ms) query to confirm the trace landed, and Ingest Key creation and its single dataset-creation permission are documented separately at /configure/environments/manage-api-keys. One page documents the full end-to-end path: signup, team creation, API key retrieval, OTel instrumentation in seven languages, and data verification in the UI.

## Find the exact limits

**PARTIAL**

PARTIAL consensus from 2 PARTIAL, 1 FAIL.

The docs define headers and quota accounting, but provide no numeric value or scope for the referenced general API limit. The rate-limit headers and their format are fully specified and the usage-versus-quota arithmetic is exact (EPM divided by 30.4, throttling accepts one of every ten events), but the general API rate limit this sentence refers to is never given a number or a scope anywhere in the docs; only Create Query Result at 10 requests per minute and its relational-fields variant at 1 request per minute are quantified. The docs describe the rate-limit header mechanism and two endpoint-specific limits but never state the general API rate limit's numeric value.

## Recover from a 429

**PARTIAL**

PARTIAL consensus from 1 PASS, 2 PARTIAL.

The central guide covers cause, formats, exceptions, and backoff, but the Events reference specifies a legacy one-field body. The recovery path is otherwise complete and unusually precise, naming the cause, both body formats, the Retry-After timestamp form, and the exact endpoints that omit that header, but the Rate Limits page says V1 endpoints return an RFC7807 Problem Detail while the V1 Create an Event endpoint documents its 429 body as the bare object {"error": "Request dropped due to rate limiting."}, so an agent writing one parser cannot trust either shape. The cause is named, both V1 and V2 response formats are described, Retry-After behavior is documented, and a recovery path (exponential backoff + Retry-After) is given.

## Verify a webhook

**PASS**

PASS consensus from 2 PASS, 1 PARTIAL.

One guide covers recipient creation, Trigger attachment, delivery, and matching X-Honeycomb-Webhook-Token against the Shared Secret. Creating the recipient, attaching it to a Trigger, and verifying the sender are all documented consistently on one page, which names the header as X-Honeycomb-Webhook-Token carrying the value of the Webhook integration's Shared Secret plus the retry, 15-second timeout, and X-Honeycomb-Webhook-Delivery-ID deduplication behavior, but the default JSON body a standard webhook receives is never shown, so the receiving endpoint must be written against a guessed schema. The webhook creation UI documents an optional Shared Secret, the headers table shows that secret is returned in X-Honeycomb-Webhook-Token, and the Trigger attachment steps are complete.

## Use the Web SDK

**PARTIAL**

PARTIAL consensus from 3 PARTIAL.

The SDK example is minimal, but endpoint forms disagree and package verification drops io from the installed package name. A minimal runnable HoneycombWebSDK example exists and the API form matches the Web quick start, but two contradictions block a clean run: the verification step names the package @honeycomb/opentelemetry-web while the install command two paragraphs above installs @honeycombio/opentelemetry-web, and the two pages give different EU endpoint values for the same option. The page provides a minimal working example but the package name contradicts itself: text says @honeycomb/opentelemetry-web while every code example imports from @honeycombio/opentelemetry-web.

## The receipt

> Some endpoints have stricter limits in addition to the general API rate limit:

The rate-limit headers and their format are fully specified and the usage-versus-quota arithmetic is exact (EPM divided by 30.4, throttling accepts one of every ten events), but the general API rate limit this sentence refers to is never given a number or a scope anywhere in the docs; only Create Query Result at 10 requests per minute and its relational-fields variant at 1 request per minute are quantified.

- [docs.honeycomb.io/api/rate-limit](https://docs.honeycomb.io/api/rate-limit)

## Agent surface notes

JSON-RPC initialize returned protocol 2025-06-18 with serverInfo name "Honeycomb Docs" version 1.0.0 and search/retrieval tool capabilities. A separate product MCP at https://mcp.honeycomb.io/mcp answers initialize with HTTP 401 and a Bearer OAuth challenge naming resource metadata at mcp.honeycomb.io/.well-known/oauth-protected-resource.

After hydration the docs header shows an Ask AI button (kapa-ask-ai-btn); one click opened the Kapa modal with an Ask a question... input, loaded from widget.kapa.ai.

## 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 3 of five tasks. Quotes shown here were re-fetched and confirmed verbatim on 2026-09-27.

Methodology: https://docsforagents.com/methodology/

Canonical URL: https://docsforagents.com/reports/honeycomb-docs-ai-agent-readiness/