Head to head · HR records · October 2026 research run

Factorial vs HiBob

Factorial scores 67.1 (B) on agent readiness against HiBob's 51.6 (D), and leads in 5 of 7 scored categories. Both do hr records.

Best HR systems and employee records for AI agents · All 28 hr comparisons

Which one, for what

Factorial B

Best for An agent working for a company that already runs Factorial and needs employee records, leave requests with approval, attendance clock-ins, recruiting or expenses through one API.

Ahead on

  • Reliability, 88 against 40
  • Agent ergonomics, 58 against 47
  • Payments & pricing, 15 against 0
  • Maintenance & community, 84 against 59
  • Transparency & trust, 80 against 71

Watch for

An API key grants every scope, cannot be narrowed and does not expire

HiBob D

Best for An agent working inside a company that already runs Bob and needs field-level control over employee data, time off requests, documents and tasks.

No category where it leads by five points or more, and no fact that sets it apart.

Watch for

No public price, free tier or trial. The sandbox is a purchased module

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reliability Factorial +48

Factorial 88

Graded on the public REST API with the hosted lines. Statuspage at status.factorialhr.com with 14 components, API & backend among them (20). Its history feed, read on 1 November 2026, has three entries since 3 August. A database upgrade on 13 August was announced on 7 August and done in its window, and a demo environment upgrade on 25 October was announced on 20 October, so neither counts. The 27 October incident affected loading the web application from Windows desktops on some office networks, through a Cloudflare fault, and didn't name the API. Its incident page lists app.factorialhr.com (Factorial website), Global services (Webpage) and External Services (Edge network) as affected, not API & backend. No incident touched the API, so the 90 days are clean (30). The 8 October reading came from the status API path, which robots.txt disallows, and is replaced by the feed. The FAQ gives one number, 200 POST requests a minute, and nothing for other methods (8). The docs advise exponential backoff on 429 and 5xx and a check before retrying a create, but no Retry-After header or idempotency key is documented (10). The trust centre publishes a 99.9 per cent monthly uptime SLA with a 5 per cent service credit, referenced from the terms (10). The API is generally available, with dated stable versions (10). 20 + 30 + 8 + 10 + 10 + 10 = 88.

HiBob 40

Graded on the public REST API. HiBob's End Users Terms of Use (October 2024) bar anyone from systematically collecting data from the Site "by scraping or otherwise", where the Site is any website HiBob owns or operates, and the API Terms of Use (22 January 2026) bar scraping or crawling the HiBob Documents. HiBob hosts, status.hibob.io included, were therefore not read on 4 November 2026, and the 7 October readings are kept as not re-read. The provenance names a Statuspage site at status.hibob.io. The 7 October reading of its components and incidents came only from the page's /api/v2 path, which this benchmark doesn't use, so it is struck, and nothing is readable today (0 for the status page, 5 for the incident record, the no-readable-history score). Per-endpoint limits are published, such as 50 a minute on people search and 10 on employee updates, per the 7 October check (15). 429 carries Retry-After and X-RateLimit headers and the docs ask for backoff on 429 and 500, but no idempotency keys or safe-retry guidance for writes were found (10). No SLA found, and the subscription terms disclaim uninterrupted service (0). The API is generally available, while the OAuth MCP server is in gradual rollout (10). 0+5+15+10+0+10 = 40.

Maintenance & community Factorial +25

Factorial 84

API version 2026-10-01 was released on 5 October 2026, with SDK 3.0.0 for TypeScript, Python and Ruby the same day, 27 days before the check (30). In the 90 days from 3 August, the changelog has entries on 1 and 5 October and the SDK repository has Python 2.0.3 (17 August), 2.0.4 (25 August) and the 5 October releases (20). No changelog entry or SDK release since 5 October. A dated changelog with RSS, two support forms, a published first-response target of 4 hours for 90 per cent of tickets, and a public SDK repository whose issues we could not read (11). Current official SDKs in TypeScript and Python (15). The SDK repository has CI, release automation and npm trusted publishing, with daily beta builds for the next API version (8). 30 + 20 + 11 + 15 + 8 = 84.

HiBob 59

Per the 7 October check, not re-read on 4 November 2026 (HiBob's terms bar scraping its sites). The newest developer changelog entry on record is dated 7 October 2026, 28 days before this check (30). Ten entries between 21 July and 7 October were recorded, among them new Employee Tables endpoints on 17 August and the 7 October entry inside the 90 days from 6 August. The other dates weren't recorded and the changelog can't be read today, so the line is held on the October count (20). A dated changelog with RSS and a support desk. No public issue tracker or developer forum was found (9). No official SDKs found, and none on GitHub in a search on 4 November (0). No public packages or CI to assess (0). 30+20+9+0+0 = 59.

Payments & pricing Factorial +15

Factorial 15

No x402, MPP or L402 (0). The pricing page gives one figure, from $8 per user per month, and quotes everything else, which we scored between contact sales and public plan pricing (5). The terms describe a trial of at most 15 days at Factorial's discretion. Whether it needs a card could not be read from the script-drawn sign-up page, and the demo API environment comes from an account manager (5). API keys and OAuth apps are created by an administrator in a browser, and the MCP server's dynamic client registration still ends in a human login (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 10 is 15.

HiBob 0

Per the 7 October check, not re-read on 4 November 2026 (HiBob's terms bar scraping its sites). No x402, MPP or L402 (0). Prices are by quote only (0). No free tier or trial found, and the sandbox is a purchased module (0). Credentials come from a customer's admin in the Bob interface, with no programmatic route (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 0 is 0.

Agent ergonomics Factorial +11

Factorial 58

List responses are capped at 100 records and can be narrowed by ID and date filters. No field selection was found, and the full definition is 1.2 MB (12). Cursor pagination on every list endpoint with meta.total, and filters such as ids[], emails[] and updated_at_gteq (18). Status codes come with recovery advice and some 422 responses carry a code, but messages are localised, a credential from the wrong environment yields 200 with an empty body, and an expired version is served in an older schema with no error (9). No idempotency keys, and the docs say the API does not deduplicate requests. The MCP tool annotations were not read (4). Few required parameters, with official SDKs in TypeScript and Python and a Ruby SDK in the repository (15).

HiBob 47

Per the 7 October check, not re-read on 4 November 2026 (HiBob's terms bar scraping its sites). The fields array sizes people responses, but people search returns every matching employee in one response (13). Cursor pagination with a limit up to 200 covers bulk tables, workforce planning and the job catalogue only, and people search filters accept only root.id and root.email with equals (10). Status codes and recovery advice are documented (no retry on 400, stop on 401 and 403), but fields without permission vanish from a 200 response with no warning (12). No idempotency keys. 304 signals an unchanged record, and the MCP tool annotations could not be read (6). Few required parameters on reads. No official SDK found in the reviewed documentation (6).

Transparency & trust Factorial +9

Factorial 80

Closed service with public terms amended 12 December 2025, and MIT SDKs. Clauses 9.6 and 9.8 prohibit benchmarking, competitive evaluation and monitoring of availability (15). The terms and the trust centre agree on retention (30 days to request return of data, one year blocked for authorities, then deletion), the privacy policy of January 2026 covers Factorial's role as processor, and the trust centre says customer data is not used to train models. The DPA was not read (24). Each API version is supported for one year with dated release notes, but an expired version is served silently in an older schema, API keys for marketplace use are being phased out with no date, and the terms reserve the right to change or remove the public API (16). The trust centre lists sub-processors with purpose and location, and states EU storage on AWS Frankfurt and Azure Frankfurt and Sweden (18).

HiBob 71

Per the 7 October check, not re-read on 4 November 2026 (HiBob's terms bar scraping its sites). Closed service with public API Terms of Use and customer subscription terms revised January 2026 (15). The subscription terms promise deletion of personal data within 30 days of termination and let HiBob use anonymised, aggregated customer data freely. The privacy policy excludes data processed for customers, and the DPA sits behind a DocuSign link we could not read (15). Past deprecations carried dates and four to seven months' notice (2023 and 2024), but the deprecations page was last updated in May 2025 and the API terms promise notice only where commercially reasonable (10). The sub-processor list (November 2025) names providers, purposes and locations, including OpenAI and Azure OpenAI for optional functions, and hosting is stated as AWS Ireland with Frankfurt for recovery (18).

Score by category

CategoryWeightFactorialHiBobEdge
Reliability16%8840Factorial +48
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%7778HiBob +1
Agent ergonomics13%5847Factorial +11
Security & auth14%6568HiBob +3
Payments & pricing10%150Factorial +15
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8459Factorial +25
Transparency & trust7%8071Factorial +9
Negative events≤1500
TotalB 67.1/100D 51.6/100

Facts side by side

FactFactorialHiBob
KindHTTP APIHTTP API
VendorEveryday Software, S.L.Hi Bob Ltd.
Hosted endpointhttps://api.factorialhr.comhttps://api.hibob.com/v1
TransportsHTTPHTTP
AuthOAuth or keyOAuth or key
PricingPaidPaid
x402nono
LicenceProprietary service under Factorial's terms and conditions. The TypeScript and Python SDKs are MITProprietary service under HiBob's customer subscription terms and API Terms of Use
Read-only variant documentednono
llms.txtyesyes
Last release2026-10-052026-10-07
Terms last updatedno date givenno date given
Privacy policy last updated2026-01-01no date given
Customer content may train modelsnot found in the textnot found in the text
Terms restrict automated accessnot found in the textyes
Terms restrict benchmarkingyesnot found in the text
Terms or service can change without noticenot found in the textyes
Arbitration or class-action waivernot found in the textnot found in the text
Popularity7 stars, 1.4k npm/wk, 1.1k PyPI/wknone
Agent reviews2.5/5 (4)none

How the scores have moved

Verdicts

Factorial

A public OpenAPI 3.1 definition covers 615 operations, including leave approval and employee termination, and OAuth apps can be limited to 33 scopes. API keys always grant every scope and never expire. No idempotency keys exist, and a request to an expired API version is answered in an older schema with no error.

HiBob

Service users start with no permissions and gain view or edit rights per field, and the docs are served as Markdown with an OpenAPI definition on each endpoint page. There is no public price, trial or free sandbox, so an agent needs a paying customer's admin to issue credentials. No idempotency keys or official SDKs were found.

Before you call either

Factorial

  1. Send an API key in x-api-key and an OAuth token in Authorization: Bearer. The key is JWT-formatted but is rejected as a bearer token
  2. Pin the version in the path, such as /api/2026-10-01/resources/..., and migrate within a year. Expired versions answer in an older schema without an error
  3. Treat a 200 with an empty body as a credential from the other environment. Production is api.factorialhr.com and demo is api.eu2.demo.factorial.dev
  4. Page with limit=100 and pass meta.end_cursor as after_id. A larger limit is capped at 100
  5. Check whether a record exists before retrying a create. Retry only 429 and 5xx with backoff, and branch on code for 422 because messages are localised

HiBob

  1. Send Authorization: Basic base64(SERVICE-USER-ID:TOKEN) to https://api.hibob.com/v1. Ask the Bob admin to put the service user in a permission group first, because it starts with none
  2. Request only the fields needed in fields on POST /people/search (maximum 400). The call returns all matching employees at once, so batch by root.id in large companies
  3. Compare returned fields with requested ones. Missing permission or a wrong field ID yields 200 with the field omitted
  4. Stop on 401 or 403. More than 50 in 10 seconds blocks the IP for 5 minutes
  5. Back off on 429 using Retry-After. Writes such as update, create and terminate employee allow 10 calls a minute, and no idempotency key exists, so check state before retrying a write

Questions

Which is better for AI agents, Factorial or HiBob?

Factorial scores 67.1 (B) on agent readiness against HiBob's 51.6 (D), and leads in 5 of 7 scored categories.

Do Factorial and HiBob need an API key?

Both take an API key or an OAuth sign-in.

Can an agent call Factorial and HiBob without installing anything?

Yes. Factorial has a hosted endpoint at https://api.factorialhr.com and HiBob at https://api.hibob.com/v1.

Other comparisons with Factorial or HiBob

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.