Orb

# Fix list: Orb

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/orb, the October 2026 research run, assessed 9 October 2026. Grade B, 66.5 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Orb: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Payments & pricing, 5 out of 100, up to 11.9 more on the total

Why it scored 5: Payment platforms take the highest step that applies on the 40-point protocol line. No x402, MPP or L402 was found in the API description, the docs index or the pricing page. Orb can collect an invoice with a Stripe Shared Payment Token that a buyer's agent granted, on request and through the merchant's Stripe account. That is a Stripe credential and not one of the three protocols, so we scored it 0 and flag the judgement (0). Core, Advanced and Enterprise all show custom pricing (0). The terms describe a 30-day free trial that Orb may grant and the rate-limit table has a trial column, but the pricing page routes every start through sales and no self-serve sign-up was found (5 of 20). A person contacts sales and creates the key in the web app (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 2. Security & auth, 55 out of 100, up to 7.9 more on the total

Why it scored 55: API keys are Bearer tokens created in the web app, separately for test mode and live mode, and can be revoked. The readiness guide says to scope and rotate keys, and no scope list or read-only key was found in the docs or the API description (20 of 30). Dashboard users have Admin, Editor or Viewer roles, and the `Dry-Run` header and preview calls let a caller check a write before committing it. Nothing limits what an API key can do (8 of 20). The API returns customer-supplied names, metadata and event properties, and no prompt-injection guidance was found (5 of 15). Audit logs name the API key or user, the targets, the IP address and the user agent and can stream to a SIEM, but they are in private preview on the Enterprise plan only (8 of 15). SOC 2 Type 2 and SOC 1 Type 2 per the trust centre, an annual third-party penetration test and a written disclosure policy with a reporting address. security.txt returned 404 and no bounty was found (14 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 3. Reliability, 77 out of 100, up to 4.6 more on the total

Why it scored 77: Hosted lines. status.withorb.com runs on Atlassian Statuspage with eight components and 90-day uptime bars (20). The history page covers August to October 2026 and lists four incidents, all in September. Three are rated minor (delayed usage-based invoices on 1 and 2 September, dashboard reporting errors on 16 September, elevated latency on 24 September) and one unrated (15 minutes of elevated latency on shared clusters on 25 September). July was not on the page read (20 of 30). Rate limits are published per category for live, test and trial accounts. The ingest reference still says no hard rate limit is enforced, which contradicts the table (14 of 15). The docs advise exponential backoff on 429 with no `Retry-After` documented, and `Idempotency-Key` on POST and PATCH makes retries safe for 48 hours (13 of 15). The pricing page lists SLAs for Enterprise only, inside the customer's agreement, and the public terms promise reasonable efforts with no figure (0). The v1 API is generally available (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 4. Transparency & trust, 65 out of 100, up to 3.1 more on the total

Made of editorial 45, provenance 84.

Why it scored 65: The platform is closed under published Platform Terms and Conditions, and the SDKs are Apache-2.0. The terms page shows no version date (17 of 30). The privacy policy of 7 May 2025 covers product data, the terms allow a data export within 30 days of termination before deletion, and the security page states seven-day snapshot retention and deletion on request. The terms let Orb use aggregated, de-identified customer data to improve the service. The DPA is available only on request (16 of 30). No deprecation policy was found. The API description marks 235 items deprecated with no removal dates, and the SDK changelog shows the ingestion `debug` parameter deprecated on 15 July and removed on 12 September 2026 (6 of 20). The security page names AWS as host and says Orb keeps a sub-processor list, which is not public. The privacy policy lists provider categories and names Stripe. No hosting region was found (6 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: withorb.com, registered 2019-10-12 (6 years) (11 of 15)
- Terms of service: read, states 5 of the 7 things a reader expects (8.3 of 10)
- Privacy policy: read, states 7 of the 8 things a reader expects (9.3 of 10)
- security.txt: not found (0 of 10)

## 5. Agent ergonomics, 84 out of 100, up to 2.6 more on the total

Why it scored 84: Graded on the REST API. List calls take `limit` (default 20, maximum 100) and a few view switches such as `include_zero_quantity_line_items`. No field selection was found, and invoice and subscription objects are large (14 of 25). Cursor pagination with `has_more` and `next_cursor`, and filters on status, customer, dates and amounts with comparison suffixes (18 of 20). Errors carry `status`, `title`, a stable `type` URI and `detail`, with a `validation_errors` array on 400 (18 of 20). `Idempotency-Key` with replay and transient-error headers, per-event idempotency keys and a `Dry-Run` header that previews most writes (20 of 20). Creating a customer needs only `name` and `email`, external IDs are accepted as aliases, and official SDKs cover six languages (14 of 15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 6. Schema & documentation, 88 out of 100, up to 2 more on the total

Why it scored 88: A public OpenAPI 3.1 description with 147 operations, 729 schemas and 79 webhook events (25). llms.txt and a Markdown twin for every docs page (10). 129 of 147 operations carry a description longer than 80 characters, and the ingest, ledger and subscription entries explain when to use each call. Few say when not to (17 of 20). The description has 688 enums and required lists, with free-form maps for `metadata` and event `properties` (13 of 15). Examples appear in about 350 places, and errors follow RFC 7807 with nine documented types and stable `type` URIs. The URIs point at an older docs path (13 of 15). The path is versioned v1 and the description says 1.0, with no dated API versions. API changes are recorded in the SDK changelogs, and the product changelog has 20 entries from October 2025 to August 2026 (10 of 15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 7. Maintenance & community, 82 out of 100, up to 1.6 more on the total

Why it scored 82: The Python SDK 4.83.0, TypeScript SDK 6.27.0 and Go SDK v1.126.0 were all tagged on 9 October 2026, the day of the check (30). The Python repository has 30 tags since 11 July 2026 (20). The product changelog's latest entry is dated 7 August 2026, with two entries in the last 90 days. A support site is linked, the terms name email and Slack support, and the SDK issue queues (four and six open items) were not read (9 of 15). Current official SDKs in six languages. The TypeScript SDK moved to a new repository and major version, with a migration guide (15). The repositories carry CI workflows with lint, build and test jobs, lockfiles and automated releases. CI results were not read (8 of 10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- unchecked: the DPA, SOC 1 and SOC 2 reports, the penetration test report and the sub-processor list. The trust centre at security.withorb.com lists them behind an access request
- unchecked: whether API keys can be scoped or made read-only. The readiness guide says to scope keys, and no page or schema describing scopes was found
- unchecked: status history before August 2026. The history page read shows three months, and the status site's robots.txt disallows `/api/`, so the incident feed was not requested
- unchecked: `https://api.withorb.com/spec.json`, the address the SDK page gives for the OpenAPI file. The host answered 403 to its robots.txt request, so nothing else was asked of it. The copy linked from the docs index was read once
- unchecked: demo.withorb.com, the demo environment the docs and pricing page link, and app.withorb.com. Whether the demo needs an account was not established
- unchecked: how a trial is granted and whether it needs a card. The terms say Orb may grant a 30-day trial, and the pricing page has only Contact Sales buttons
- unchecked: PyPI download counts (PyPI's robots.txt closes the path), CI results and the SDK issue threads
- unchecked: the effective date of the Platform Terms and Conditions. The page shows a Version label with no value in the text we read
- unchecked: whether the API still accepts the `debug` parameter on ingestion. The SDK changelog records its removal on 12 September 2026 and the reliability page still mentions `debug=True`. No deduction was taken because the SDK changelog marked it deprecated first
- No MCP server was found in the docs index, the home page or the SDK repositories. The earlier Node SDK changelog has MCP fixes dated December 2025, and neither TypeScript repository contains an MCP package today
- Neither the Platform Terms nor the website terms of use (effective 21 April 2025) were found to forbid automated access or benchmarking
- changelog.withorb.com has no robots.txt (the address returns the changelog page). The entries were read from the RSS feed the page links
- The lead listed SDKs for Python, Node, Ruby and Go. The docs also list Java and Kotlin. The lead's other points held

## Weaknesses

- All three plans (Core, Advanced, Enterprise) show custom pricing only, and the pricing page sends new customers to sales
- No list of API key scopes or a read-only key was found. Keys are created in the web app and revocation is irreversible
- Audit logs are in private preview and limited to the Enterprise plan
- The DPA and SOC reports sit behind an access request, and no public sub-processor list or hosting region was found
- The SDK changelog shows the ingestion `debug` parameter removed on 12 September 2026 while the reliability page still refers to it, and the product changelog has no entry

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Call `https://api.withorb.com/v1` with `Authorization: Bearer <key>`. A person creates the key in the Orb web app under organisation settings, in test mode or live mode
- Send an `Idempotency-Key` of at most 64 characters on every POST and PATCH. A 409 means the first request is still running or the body changed
- Add `Dry-Run: True` to check a write first, and confirm `Orb-Dry-Run: true` in the response. Ingest, amend and deprecate event calls ignore it and run for real
- Do not send `Idempotency-Key` with `Dry-Run`. Orb returns a validation error
- Page with `limit` (default 20, maximum 100) and pass `next_cursor` back as `cursor` while `has_more` is true
- Back off exponentially on 429. Test mode allows 5 writes a second and 2,000 ingested events a minute

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.