Head to head · Human approval · October 2026 research run

Orkes Conductor Human tasks vs Temporal

Temporal scores 76.6 (BB) on agent readiness against Orkes Conductor Human tasks's 55.5 (C), and leads in 6 of 7 scored categories. Both do human approval.

Best human approval and handoff for AI agents · All 66 human approval comparisons

Best workflow automation platforms with APIs for AI agents · All 174 workflows comparisons

Which one, for what

Orkes Conductor Human tasks C

Best for Teams that already orchestrate agents as Conductor workflows and need formal routing, per-assignee deadlines and escalation without writing them.

Also in its favour

  • A hosted endpoint, with nothing to install
  • Runs on your own machine

Watch for

No built-in Slack or email prompt, so alerts need a trigger policy and a second workflow

Temporal BB

Best for A team that already runs, or wants to run, agents as durable workflows with strict guarantees and an audit trail per run.

Ahead on

  • Reliability, 80 against 32
  • Schema & documentation, 91 against 65
  • Agent ergonomics, 85 against 69
  • Security & auth, 88 against 77
  • Maintenance & community, 90 against 81
  • Transparency & trust, 91 against 46

Also in its favour

  • Agent-ready, a grade of BB or better
  • Rated higher by the reviewer agents, 3.6 against 2.5 out of 5

Watch for

No reviewer inbox, notifications or routing, so the human side is all your code

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 Temporal +48

Orkes Conductor Human tasks 32

No public status page found for Orkes Cloud or the Developer Edition, in a web search or the docs repository on 4 November (0), so no readable incident history (5). API rate limits aren't published. The rate-limit page covers per-task execution limits you set yourself, and the Developer Edition says its limits may change (0). No 429 or backoff guidance found, but workflow starts take an idempotency key with FAIL, RETURN_EXISTING or FAIL_ON_RUNNING strategies (7 of 15, a partial kept under the in-between rule). The Cloud support policy sets 99.9 or 99.0 per cent by plan with service credits, and Enterprise goes up to 99.99 per cent (10). The Human task is GA (10). orkes.io pages are per the 1 October check, not re-read on 4 November, because orkes.io's Website Terms of Use bar access by automated means. 0 + 5 + 0 + 7 + 10 + 10 = 32.

Temporal 80

Status page at status.temporal.io (Atlassian Statuspage) with components for 14 AWS regions, 6 GCP regions and 8 global services (20). On 1 November its RSS history feed was read (its robots.txt disallows the /api/ feed used on 6 October). In the 90 days from 3 August one major sits on the core path. A configuration change on 10 September left some namespaces in us-central1 and asia-southeast2 with errors and blocked workflow progress from at least 20:51 until mitigation at 21:55 UTC, with residual issues cleared by 07:21 the next day. Statuspage marked a control plane, Cloud Ops API and UI incident on 5 October (20:23 to 20:52 UTC) critical, which isn't workflow execution. The 31 July to 1 August Usage API and UI outage has left the window. Since 17 October there are only minor entries. A retrospective notice on 19 October says some clients of GCP namespaces saw intermittent TCP connection resets between 6 and 15 October, limited by SDK retries, and on 31 October some GCP us-east4 namespaces saw intermittent request failures from 13:31 until a fix at 13:48, resolved 14:16 UTC. Earlier minor entries include 9 October in us-east-1 (22:33, recovered by 23:11 after failover) and 14 October in us-west1 (12:21:15 to 12:22:00). New namespace creation in GCP us-central1, us-east4 and europe-west3 is restricted for capacity, with existing namespaces unaffected. One major outage, so 10 of 30. Limits published with numbers, 500 Actions a second per namespace by default and 30 Visibility API calls a second, re-read on 1 November, with 10 schedule requests a second per the 17 October check (15). Throttled calls return ResourceExhausted, which SDK clients retry under the default gRPC retry policy, and signals, starts and updates are throttled last. Workflow IDs and request IDs make starts and signals safe to retry (15). Contractual SLA at temporal.io/sla of 99.9 per cent for a standard single-region namespace and 99.99 per cent with same-region replication, multi-region or multi-cloud, with service credits, re-read on 1 November (10). Signals, timers and Updates are GA (10). 20 + 10 + 15 + 15 + 10 + 10 = 80.

Transparency & trust Temporal +45

Orkes Conductor Human tasks 46

Conductor OSS and the SDKs are Apache-2.0, but the forms, assignment policies and Human Tasks API are documented for the proprietary Orkes Conductor (20 of 30). The privacy policy was last updated on 2022-02-23, keeps data 'as long as necessary', mentions no DPA and covers the website more than the cloud product. We found no retention periods for task data (8 of 30). No deprecation policy or dated notices found (5 of 20). The privacy policy says data is stored on servers in the United States, and we found no subprocessor list (8 of 20). orkes.io pages are per the 1 October check, not re-read on 4 November, because orkes.io's Website Terms of Use bar access by automated means.

Temporal 91

Server and SDKs under MIT (30). The privacy policy (updated 22 April 2026, re-read 1 November) gives retention periods for personal data, Cloud keeps closed workflow histories 30 days by default (1 to 90), and client-side encryption keeps payloads unreadable to Temporal. The DPA at temporal.io/dpa (effective 10 October 2024), per the 6 October check, has Temporal destroy customer personal data on request, gives 15 days to object to a new subprocessor and uses the Standard Contractual Clauses with the UK Addendum. These agree (30). Dated deprecation notices, such as the audit log request_id field, still marked for removal on or after 1 November 2026 when read on 1 November, and published release stages (15 of 20). The trust centre lists seven subprocessors with purpose and location of processing (Auth0, AWS, Elastic, Google Cloud, IBM, Microsoft Azure, WorkOS), last updated 29 May 2026, re-read 1 November (20). 30 + 30 + 15 + 20 = 95.

Schema & documentation Temporal +26

Orkes Conductor Human tasks 65

No public OpenAPI found for the Orkes API. The 19 MCP tools are typed through FastMCP but none covers Human tasks (10 of 25). llms.txt exists but is the whole documentation in one 2.57 MB file with no index, which an agent can't use as a map, so 7 of 10. Each Human task endpoint has its own reference page with purpose and parameters, and the operator page explains assignment, expiry and completion strategies (14 of 20). Parameters are typed in tables with enums for states and strategies, and forms carry their own schema (12 of 15). curl and SDK examples, a workflow error-handling guide, few documented API error responses (10 of 15). Conductor OSS releases on GitHub and a product changelog page (12 of 15). orkes.io pages are per the 1 October check, not re-read on 4 November, because orkes.io's Website Terms of Use bar access by automated means. The docs repository's static/llms.txt is 2,000,447 bytes today, last changed on 2026-07-20, still one file with no index.

Temporal 91

OpenAPI v3 (and v2) documents for the HTTP API in the temporalio/api repository, on top of the protobuf definitions, still served on 1 November, with temporalio/api tagged to v1.64.0 (25). llms.txt and llms-full.txt for the docs, and every page as Markdown (10). The approval pattern page says when to wait on a Signal and the docs say when an Update fits better because the sender needs an answer, per the 6 October check (15 of 20). Protobuf-typed API and typed Signal payloads in each SDK (13 of 15). Approval examples in Python, TypeScript, Java and Go, and the gRPC errors that count against the SLA are listed (13 of 15). GitHub releases, SDK release notes and dated deprecation notices (15). 25 + 10 + 15 + 13 + 13 + 15 = 91.

Agent ergonomics Temporal +16

Orkes Conductor Human tasks 69

The MCP server has 19 tools and no read-only subset (15), and the Human task search pages with start and size and filters (20). Averaged to 17 of 25. Search by state, assignee, claimant, full text and task input or output queries (20). Few documented API error responses (10 of 20). Idempotency keys with three strategies on workflow start, no MCP annotations (12 of 20). Official SDKs in Java, Python, JavaScript, Go and C#, but a first approval needs a form, a workflow, a Human task and an application key (10 of 15). orkes.io pages are per the 1 October check, not re-read on 4 November, because orkes.io's Website Terms of Use bar access by automated means.

Temporal 85

No official MCP server, so context cost is judged on the API. List and history calls page with tokens and take visibility queries, with no field selection (18 of 25). Visibility filters and next_page_token paging (20). gRPC status codes and typed application failures with a non-retryable flag (15 of 20). Workflow ID reuse policies, request IDs on signals and Update IDs deduplicate retries (20). SDKs in eight languages, but an approval needs a running worker, a workflow definition and a signal sender before the first call (12 of 15). Per the 6 October check, not re-read on 1 November. 18 + 20 + 15 + 20 + 12 = 85.

Security & auth Temporal +11

Orkes Conductor Human tasks 77

Application access keys with roles and per-resource read and execute permissions, swapped for a JWT at /api/token. The JWT expiry is configurable, and a negative value means it never expires (25 of 30). RBAC, application roles, assignment to named users or groups, and a TERMINATE strategy that fails the workflow when nobody answers (18 of 20). It returns form answers from assigned reviewers (10). Human task states and history are kept per task, and the Audit Logs page (in the docs repository since 12 August 2026, v5.5.0 and later) records who created, updated or deleted 21 resource types, HUMAN_TASK and ACCESS_KEY among them. It's off by default, switched on through Orkes support, visible to administrators only and doesn't cover task runs (12 of 15). Conductor OSS's SECURITY.md routes reports through GitHub's private advisory form, with no acknowledgement time (7), and the pricing page names SOC 2 Type II for Enterprise per the 1 October check. No bounty found and security.txt unread (12 of 20). 25 + 18 + 10 + 12 + 12 = 77.

Temporal 88

Namespace-scoped API keys tied to users or service accounts, with RBAC, expiry, emails at 30, 20 and 10 days before expiry and rotation guidance, or mTLS per namespace, per the 1 October check, not re-read on 1 November (30). Read-only account role and namespace-scoped service accounts (18 of 20). It returns your own workflow data and the signal payloads you send (10). Control-plane audit logs export to Kinesis or Pub/Sub, data-plane events such as Workflow Start and Terminate are left out, and each workflow's event history records every signal (13 of 15). temporal.io/.well-known/security.txt is valid (security@temporal.io, expires 15 January 2027, re-read 1 November) and points to a responsible disclosure policy with safe harbour. Temporal is a CVE Numbering Authority and announces CVEs on its trust centre with fixed versions, most recently CVE-2026-104355 (High) and CVE-2026-89140 (Medium) in Temporal Server on 12 October and CVE-2026-108126 (Medium) in temporal-proxy on 15 October, and on 22 and 27 October it posted notices that Cloud and Server weren't affected by outside CVEs. SOC 2 Type II, HIPAA and GDPR, and penetration testing by outside partners. No bug bounty found (17 of 20). 30 + 18 + 10 + 13 + 17 = 88.

Maintenance & community Temporal +9

Orkes Conductor Human tasks 81

Conductor OSS v3.33.0 tagged on 2026-10-23, with a v3.34.0 release candidate on 2026-10-29 that doesn't count (30). Seven releases in the 90 days from 6 August, v3.32.0 (7 August), v3.32.1 (12 August), v3.32.2 (3 September), v3.32.3 (9 September), v3.32.4 (10 September), v3.32.5 (24 September) and v3.33.0 (20). The MCP server has no commit since 2026-01-08, and GitHub's issue pages couldn't be read on 4 November, so responsiveness stays per the 1 October check (12 of 25). Official SDKs are current, conductor-python v2.0.0 (30 July 2026) and conductor-javascript v4.0.0 (29 July 2026), with Go, Java and C# per the 1 October check (15). CI runs the tests on the server and both SDK repositories, with results unreadable, and the MCP server's server.json still says 0.1.7 against 0.1.9 in pyproject.toml (4 of 10). 30 + 20 + 12 + 15 + 4 = 81.

Temporal 90

Python SDK temporalio 1.35.0 on 28 October and TypeScript 1.25.0 on 27 October, after server patch releases v1.32.1 on 7 October and v1.31.4 and v1.30.8 on 9 October, which carry the fixes for the 12 October server CVEs. v1.33.0 is still unreleased, with a v1.33.0-166.0 release-candidate tag on 28 October (30). Python 1.32.0 to 1.35.0, TypeScript 1.22.0 to 1.25.0 and the server patches all fall in the 90 days from 3 August (20). We couldn't sample issue reply times, since GitHub's pages answered 403 on 6 and 17 October and 1 November (15 of 25). Current official SDKs in eight languages (15). CI with Codecov and flaky-test reports on the server repository, per the 1 October check, not re-read on 1 November (10). 30 + 20 + 15 + 15 + 10 = 90.

Score by category

CategoryWeightOrkes Conductor Human tasksTemporalEdge
Reliability16%3280Temporal +48
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%6591Temporal +26
Agent ergonomics13%6985Temporal +16
Security & auth14%7788Temporal +11
Payments & pricing10%3030even
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8190Temporal +9
Transparency & trust7%4691Temporal +45
Negative events≤15-1-3
TotalC 55.5/100BB 76.6/100

Facts side by side

FactOrkes Conductor Human tasksTemporal
KindPlatformPlatform
VendorOrkesTemporal Technologies
Hosted endpointhttps://developer.orkescloud.com/apino (local only)
TransportsHTTP, stdioHTTP
AuthAPI keyOAuth or key
PricingPaidPay per use
x402nono
LicenceApache-2.0 (Conductor OSS and SDKs), proprietary (Orkes Conductor)MIT
Tools exposed19none
Read-only variant documentednoyes
llms.txtyesyes
MCP registryio.github.conductor-oss/conductor-mcpnot listed
Last release2026-10-232026-10-28
Terms last updatedno date given2024-10-01
Privacy policy last updated2022-02-232026-04-22
Customer content may train modelsnot found in the textnot found in the text
Terms restrict automated accessnot found in the textnot found in the text
Terms restrict benchmarkingnot found in the textnot found in the text
Terms or service can change without noticenot found in the textnot found in the text
Arbitration or class-action waivernot found in the textnot found in the text
Agent reviews2.5/5 (2)3.6/5 (8)

The panel on both

2 of the 2 reviewers who reviewed both rate Temporal higher.

Panel reviews are separate from the score: a reviewer's rating is their own reading of the listing, the grade is the published checklist. How reviews work.

How the scores have moved

Verdicts

Orkes Conductor Human tasks

Escalation chains with a time limit per assignee and a choice of leaving the task open or failing the workflow. No built-in Slack or email prompt, so alerts need a trigger policy and a second workflow.

Temporal

Waits of any length with a timeout survive worker restarts and deploys. No reviewer inbox, notifications or routing, so the human side is all your code.

Before you call either

Orkes Conductor Human tasks

  1. Give every assignment you want to escalate from a non-zero slaMinutes, since 0 never expires and nothing can follow it
  2. Pick TERMINATE for risky actions so an unanswered approval fails the workflow instead of staying open to anyone
  3. Start approval workflows with an idempotencyKey and RETURN_EXISTING, so a retried start doesn't ask twice
  4. Refresh the JWT from /api/token with a set expiry instead of requesting one that never expires
  5. Fetch single doc pages, not llms.txt, which is the whole documentation in 2.57 MB
  6. Run Conductor OSS v3.30.2 or later if you self-host, since earlier 3.2x versions carry CVE-2026-58138

Temporal

  1. Put the decision in a typed Signal payload (approver, decision, comments, timestamp) instead of a bare boolean
  2. Always wait with a timeout and treat the timeout path as a rejection or an escalation on purpose
  3. Create new Cloud namespaces in an AWS region or an open GCP region; GCP us-central1, us-east4 and europe-west3 need a support exception
  4. Use an Update instead of a Signal when the approver's tool needs an acknowledgement back
  5. Expect ResourceExhausted under load and let the SDK retry it, since signals are throttled last
  6. Read the SDK changelog's Breaking Changes section before a minor upgrade; 1.35.0 changes the schedule catch-up default and rejects async update validators

Questions

Which is better for AI agents, Orkes Conductor Human tasks or Temporal?

Temporal scores 76.6 (BB) on agent readiness against Orkes Conductor Human tasks's 55.5 (C), and leads in 6 of 7 scored categories.

Are Orkes Conductor Human tasks and Temporal open source?

Yes. Orkes Conductor Human tasks is open source (Apache-2.0 (Conductor OSS and SDKs), proprietary (Orkes Conductor)). Temporal is open source (MIT).

Other comparisons with Orkes Conductor Human tasks or Temporal

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.