Agentic Index
Humanly vs Metix AI (2026)
Both vendors last verified
Humanly and Metix AI, formerly OpenJobs AI, both run autonomous recruiting, and the split is funnel automation versus full recruiter replacement. That verdict is the Agentic Index coverage score, graded from each vendor's own published materials.
Humanly's AI Recruiter automates top of funnel engagement, screening, scheduling and interviews for high volume hiring, while Metix AI's Mira runs the full cycle from a five minute brief through sourcing, outreach and booked first interviews, with every outreach message approved before it sends. Humanly sells on contact, and Metix AI starts at $99 a month on Starter with 5,000 credits, after a free 14 day trial with 500 credits. Choose Humanly for a proven high volume play. Choose Metix AI for a more aggressive autonomy bet.
This comparison is published by Agentic Index, an independent agentic AI vendor research platform. Humanly and Metix AI are each graded against the same 14 capability Agentic Index taxonomy, from the vendor's own public materials under the Agentic Index verification standard, alongside 1,140 researched vendors. No vendor pays for placement and no vendor has reviewed this page. How this evidence is graded
Choose Humanly if
- High volume and frontline hiring funnels are your core problem.
- A proven vendor in production at scale reduces your risk.
- Top of funnel automation with human recruiters closing fits your model.
Choose Metix AI if
- You want to test full cycle autonomous recruiting from brief to schedule.
- Being early on an agent first platform fits your risk appetite.
- End to end ownership by one agent appeals over funnel handoffs.
| Feature | H Humanly |
M Metix AI |
|---|---|---|
| Action & orchestration | ||
|
Integrations & Tool Calling Ability to connect agents to real systems through native integrations, OAuth-authenticated actions, custom tools, APIs, webhooks, or MCP-compatible tools. |
||
|
HumanlyIntegrations & Tool Calling Two way integrations connect Humanly to the customer's existing ATS, job boards and HR systems, with no ripping and replacing. The nine named systems are Workday, SAP SuccessFactors, Oracle Taleo, iCIMS, Greenhouse, Avature, UKG, Dayforce and Employ. Screening data, dispositions and transcripts are written back to the applicant tracking and HRIS systems, so the hiring record stays in one place, with Humanly sitting on top of the existing system as an engagement layer. The AI Recruiter syncs with recruiter and hiring manager calendars to book interviews automatically, with no email back and forth, writing into two parties' calendars with authenticated access. Candidate reach extends across job boards and social surfaces. Every named partner sits in the recruiting stack, and Humanly publishes no connector marketplace or integration directory. Sourcehumanly.io homepage and /products/ai-recruiterread 2026-09-08 |
||
|
Metix AIIntegrations & Tool Calling Deel is the global strategic partner, connecting hiring to employer of record payroll and compliance under the line "one AI recruiter, hiring across 130+ countries". Interviews are booked onto the customer's own calendar, an authenticated write into a system Metix does not own, and Metix does not name the calendar provider. Outreach goes to candidates through messages the customer has approved. Metix names no other connected system, no connector directory, ATS list or HRIS, and no integration count. Custom integrations are sold as a line item on the $299 Scale plan without further description, so integration beyond the default is bespoke work. The developer site from before the rebrand claimed a single integration for LinkedIn, Indeed and major ATS platforms. That host now redirects to platform.metix.ai, a data API for people, jobs and companies that makes no ATS, LinkedIn or Indeed claim, so the earlier claim is unconfirmed but not disproved. Sourcemetix.ai/pricing, homepage and platform.metix.airead 2026-09-12 |
||
|
Workflow Orchestration Ability to sequence, branch, retry, route, and combine deterministic workflow nodes with autonomous agent steps. |
||
|
HumanlyWorkflow Orchestration Five stages run as one system, from first message to scheduled interview, in place of five disconnected tools. Humanly engages every candidate, screens for real fit, conducts structured interviews, schedules the next step and delivers ranked candidates with transcripts and clear reasoning. Screening and scheduling run through configurable workflows, so the customer defines the sequence. The candidate is conversed with, the recruiter receives a ranked shortlist, and the hiring manager's calendar is booked. The AI Recruiter syncs with recruiter and hiring manager calendars, so two internal parties and the applicant are coordinated in one pass. Humanly addresses recruiters, talent leaders, operations leaders and candidates separately. Sourcehumanly.io homepage and /products/ai-recruiterread 2026-09-08 |
||
|
Metix AIWorkflow Orchestration Mira, the agent, runs hiring from brief to booked interview in five named stages, which are AI role briefing, talent search, personalized outreach, qualification and scheduling. The customer sets the brief and approves each message, and Mira runs search through booking. In the published funnel, 2,847 candidates are screened, outreach goes to 312 at 11 percent, 48 reply with interest at 15 percent, and first interviews are booked across the week, with each stage narrowing on a stated criterion. The customer writes the brief in a five minute setup, naming what is essential and what can flex, and Mira confirms it. Every outreach message is approved before it sends, so the flow pauses at that point and then resumes. Three parties take part. The hiring manager writes the brief and approves, Mira plans and executes, and the candidate is contacted and qualified, with the interview landing on a calendar. Metix names no orchestration runtime or handoff between agents, and there is no builder for authoring a different flow. The sequence is Metix's own recruiting process, with the brief as its parameter. Sourcemetix.ai homepage and /pricingread 2026-09-12 |
||
|
Triggers & Channel Coverage How agents wake up and where they work: schedules, webhooks, message events, CRM events, inbox events, chat, email, voice, and collaboration tools. |
||
|
HumanlyTriggers & Channel Coverage Every applicant is greeted within seconds on SMS, chat or email, with no waiting and no drop offs, so a candidate who applies at 9pm is on the recruiter's calendar the next day. Past applicants are contacted again from the CRM when a role opens, and configurable workflows evaluate, score and schedule automatically. Channels cover SMS, email, web chat, voice, phone and video. Structured AI video, voice and asynchronous interviews run 24/7 on the candidate's schedule, with automated reminders and follow up. Sourcehumanly.io homepage and /products/ai-recruiterread 2026-09-08 |
||
|
Metix AITriggers & Channel Coverage Work starts when a role is defined, and candidate replies drive what happens next. The customer starts from a job description file, a similar profile, or a LinkedIn or job URL. Outreach is personalized for each candidate, and every message is approved before it sends. Replies are read and classified for real interest, and a candidate who clears the quality bar triggers both the booking and the credit. Qualified, interested candidates then show up on the customer's own calendar, with booked slots shown by day. Metix says a brief produces candidates ready to interview within 24 hours. Metix lists no outreach channels. Email is clear from context but not named, and no second outbound channel is stated. Metix names no schedule control, cadence setting or event framework reaching outside the product. The OpenJobs site from before the rebrand described continuous automated outreach and follow ups across email and WhatsApp. WhatsApp no longer appears on metix.ai and is unconfirmed, not retired. Sourcemetix.ai homepage and /pricingread 2026-09-12 |
||
| Knowledge & context | ||
|
Knowledge Grounding & RAG Ability to ground agent behavior in company data through document ingestion, retrieval, external knowledge APIs, semantic search, or RAG layers. |
||
|
HumanlyKnowledge Grounding & RAG Two maintained corpora sit under the agent. The customer's own is a Talent CRM and built in ATS that hold the organization's candidates, pipelines, transcripts and dispositions, and the platform re-engages past applicants from that store when a new role opens. The second is a network of 600M+ profiles that Humanly holds for sourcing. The agent reads both in every conversation, giving dynamic answers and job details tailored per role, and returns a ranked shortlist with transcripts behind every name. The record persists between runs, covers an entire applicant history and is queried, not assembled per job. Humanly names no retrieval structure behind the grounding and does not say whether answers cite their sources. Sourcehumanly.io homepage and /products/ai-recruiterread 2026-09-08 |
||
|
Metix AIKnowledge Grounding & RAG A maintained index holds 900 million profiles, 90 million job postings and 30 million companies, searched through a published query language. Of those job postings, 10 million are live on any given day, and the jobs index tracks daily active postings. The index is standing and refreshed on a cycle, not assembled per request, and it scales far past any context window. Retrieval is specified down to the field level. A Query Spec tree has named fields, match, equality and range operators, and all, any and not composers. A natural language route translates a sentence into the same tree, and agents read a live contract route at run time to get field names and operators, so they can query the schema instead of guessing it. All three datasets take the same query tree. In the hiring product, Metix captures the role's real requirements upfront in a brief that Mira confirms, then searches and grades candidates against that brief instead of keywords. The corpus is Metix's own, and Metix names no ATS sync or customer document store, so the customer's own hiring history is not an input. Sourceplatform.metix.ai and metix.ai homepageread 2026-09-12 |
||
|
Memory & State Persistence Ability to persist context across a run, conversation, workflow, user, team, or longer-term memory layer. |
||
|
HumanlyMemory & State Persistence No agent memory is published, whether a memory layer, cross session state store or retained preference set, and Humanly names no memory scope or lifetime. The Talent CRM and built in ATS hold candidates, pipelines, transcripts and dispositions, and past applicants are contacted again from that store. Humanly names no retention or deletion control over what the agent has learned, as distinct from candidate data, and no store of taught criteria. Configurable knockout criteria are set by the customer in advance and can be changed. Sourcehumanly.io homepage and /products/ai-recruiterread 2026-09-08 |
||
|
Metix AIMemory & State Persistence The data platform is stateless, and each role starts from a fresh brief. Integration takes three calls, which get a key, search and read the records, and a search returns encrypted identifiers that are then exchanged for records. Nothing in the published contract carries state between requests. Metix sets out rate limits, credit rules and refusal behavior in detail, but names no session, context or memory concept. In the hiring product the brief is an input for each role that the customer writes in five minutes, not a store the agent builds across roles. Metix does not say that Mira carries what it learned on one search into the next, or that a rejection on one role informs another. The index of 900 million profiles persists, but it is Metix's data product, and credits, allowances and usage history are account bookkeeping. Metix names no memory layer or retained preference set, no decision trace across runs, no scope or lifetime, and no customer control to inspect or purge agent state. Sourceplatform.metix.ai and metix.ai homepageread 2026-09-12 |
||
| Control & trust | ||
|
Human Oversight & Guardrails Approval steps, consent checkpoints, escalation rules, structured guardrails, policy constraints, and pause/resume controls. |
||
|
HumanlyHuman Oversight & Guardrails A candidate conversation can leave the agent mid flow through an instant handoff to a human chat. Recruiters can override the agent's output, and configurable knockout criteria, set in advance, are the customer's own boundary on what the agent may advance or reject. The agent engages, screens, scores and books the interview without a recruiter in the loop, so a candidate who applies at 9pm is on a calendar the next day. Humanly names no approval queue, confidence threshold, hold before an invitation is sent or way to reverse a knockout. Sourcehumanly.io/solutions/screen and homepageread 2026-09-08 |
||
|
Metix AIHuman Oversight & Guardrails Every outreach message needs the customer's approval before it is sent. The customer writes the brief and approves each message, and Mira then runs search through booking, sourcing candidates, screening them for interest and booking interviews. Nothing reaches a candidate until a person has approved that specific message, and this is how Metix separates its self serve product from its own managed service. Recruiters also review a validated shortlist before moving to conversations. A second check comes after the agent and is enforced through pricing. Candidates reach the customer only after clearing its quality bar for fit and real interest, the customer sets that bar in the brief, and if nobody clears it no credit is spent. Delivery service, where "someone on our side works the role from day one", is a separate quoted purchase and not one of the four plans. On the plans, a Metix person does not give a last confirmation on the shortlist, so the approval step belongs to the platform, not to a managed engagement. Sourcemetix.ai/pricing FAQ and homepageread 2026-09-12 |
||
|
Security, Identity & Governance RBAC, SSO, auditability, encryption, least-privilege tool access, compliance posture, and data handling policy. |
||
|
HumanlySecurity, Identity & Governance No security or trust page is published, and the footer links only a Privacy Policy and Terms and Conditions. Humanly shows no trust center, compliance page or certification badges, and names no certification, SSO or access control. Customers include Microsoft, MGM Resorts, Mutual of Omaha, Trinity Health and Reid Health. Sourcehumanly.io homepage navigation, footer and FAQread 2026-09-08 |
||
|
Metix AISecurity, Identity & Governance SOC 2 Type II and ISO 27001 are both in progress, so Metix holds no security certification yet. Every page carries a SOC 2 Type II badge and an ISO 27001 badge in its footer, each captioned In progress and linking to section 7 of the privacy policy. A GDPR badge has no such caption, and it reflects Metix's own claim of compliance, not a third party attestation. Metix names no single sign on or SCIM, no role based access or permission model and no admin console, on the hiring site or the developer platform. The nearest thing is an API key with scopes and quota shown on a status route, which authenticates a developer and does not govern who in the customer's organization may act. Metix also publishes Terms of Service, a Privacy Policy with a security section and a Master Service Agreement. The data platform is designed for data minimization, so a search returns encrypted identifiers and "nothing leaves the index until you name it". That governs what the API discloses, not who may act. Sourcemetix.ai footer, /pricing and platform.metix.airead 2026-09-12 |
||
|
Observability & Auditability Traces, logs, execution histories, metrics, audit events, and debugging detail for production agent behavior. |
||
|
HumanlyObservability & Auditability The agent's conversations are recorded as transcripts. Humanly delivers a ranked shortlist with transcripts behind every name, and ranked candidates come with transcripts and clear reasoning. The transcript shows what the agent did, and structured scoring against a rubric and knockout criteria lets a reviewer reconstruct the basis of each decision. Humanly's product descriptions name no retained action log, routing record, override capture, version history or export. Timestamps, routing decisions, recruiter overrides and audit export come up only in a Humanly marketing article, its AI recruiting tools roundup. Sourcehumanly.io homepage and /products/ai-recruiterread 2026-09-08 |
||
|
Metix AIObservability & Auditability The funnel view shows 2,847 candidates screened with 11 percent reached out, outreach sent to 312 with 15 percent replying with interest, and 48 interested candidates, alongside booked interview slots by day. Reports have their own place in the site navigation, and the developer platform offers a usage view and a free key status route that reports scopes and quota. Under the credit model each unlocked candidate is accounted for individually, so a customer can reconcile what the agent delivered against what was charged. Metix names no run log, action history or audit trail, and no export or retention period. It does not say whether it keeps a record of which candidates Mira screened out and why, which messages it sent and when, or how it reached a qualification verdict. Sourcemetix.ai homepage, /pricing and platform.metix.airead 2026-09-12 |
||
|
Deployment & Data Residency Deployment modes and options, including SaaS, dedicated cloud, VPC, on-prem, hybrid, local runtime, and self-hosting. |
||
|
HumanlyDeployment & Data Residency The product is a hosted conversational service for candidates, delivered over SMS, chat and video. Humanly names no deployment choice, whether a region, regional instance, EU or other residency option, VPC or customer tenant. No cloud provider or hosting region is published. Sourcehumanly.io homepage, navigation and footerread 2026-09-08 |
||
|
Metix AIDeployment & Data Residency Four hiring plans set candidate and role allowances and billing cadence, with a 10 percent yearly discount. No plan carries a region, residency or hosting option, including the $299 Scale plan, whose named extras are custom integrations and an SLA. The developer platform publishes rate limits, page sizes and a changelog, and names no region either. Customers cannot choose a region, and Metix does not say where data is kept. There is no self hosted, on premises or private deployment option that the customer controls. Hiring across 130+ countries through the Deel partnership describes where candidates are, not where data sits, and the Delaware registered office says nothing about infrastructure. Sourcemetix.ai/pricing, homepage and platform.metix.ai; metix.airead 2026-09-12 |
||
| Solution readiness | ||
|
Prebuilt Agents, Templates & Packs Ready-made workflows, packaged employees, templates, blueprints, industry solutions, and role-specific agents that reduce time-to-value. |
||
|
HumanlyPrebuilt Agents, Templates & Packs AI Recruiter, AI Interviewer, AI Notetaker, Talent CRM and ATS each have their own page, and Humanly sells them as one platform. Three are agents doing distinct jobs at different points of the funnel. An Interview Builder offers best practice interview templates that a team customizes for questions, depth, tone and scoring. Five solution areas (Source, Attract, Screen, Schedule, Interview) and ten industry pages sit alongside as configurations. Humanly pitches everything you need, built into one platform, as one system from first message to scheduled interview, in place of five disconnected tools. It names no way to adopt, decline or swap the three agents separately, and no browsable agent catalog, gallery or way to clone and configure an agent. The Interview Builder holds interview templates, not agents. Sourcehumanly.io/products/ai-interviews and homepage navigationread 2026-09-08 |
||
|
Metix AIPrebuilt Agents, Templates & Packs Four installable agent skills teach an assistant to use the API, and one agent, Mira, does the hiring work. The skills install with a single command, npx skills add MetixAI-Official/metix-skills, and have their own documentation page, where an All four skills link serves as the catalog. Each skill installs on its own and works in any agent, with or without MCP. The skills teach an assistant the endpoints and have it read field names and operators from a live contract route at run time. There is no browsable marketplace. The skills teach a customer's assistant how to call Metix and do not themselves source, screen or write outreach. On the hiring side the product ships one agent, Mira, and the plans differ by candidate and role allowance, not by which agents a customer adopts. The platform from before the rebrand described prebuilt agents for parsing, matching and JD generation, which no longer appear in that form. Sourceplatform.metix.ai and its skills documentation indexread 2026-09-12 |
||
| Platform extensibility | ||
|
Model Flexibility & Routing Ability to work across multiple foundation models, route tasks to different models, or let buyers bring their own providers and keys. |
||
|
HumanlyModel Flexibility & Routing No model provider, model picker, admin setting, routing description or bring your own model option is published. Humanly describes its AI's work in detail, covering conversational engagement across six channels, structured screening against rubrics and knockout criteria, and ranked shortlists with transcripts and reasoning, and it states a training base of millions of candidate interactions. It presents all of this as its own capability, with conversational quality as the difference from keyword scanning. Sourcehumanly.io homepage and /products/ai-recruiterread 2026-09-08 |
||
|
Metix AIModel Flexibility & Routing The product uses its own AI for search and matching. Metix names no LLM, model or provider, including for capturing the brief, matching and grading, writing outreach and qualifying replies. Customers have no choice of model, and Metix describes no routing, model picker, admin setting or way to bring your own key. The developer platform sets out endpoints, the query language and credit arithmetic to the quarter cent, and says contact enrichment "is announced and not open, not callable, not billable". Metix publishes an MCP server and four agent skills so a customer can drive it from Claude, Codex or any assistant. In that case the customer's assistant is calling Metix, and picking the assistant does not change the model behind Metix's matching or outreach. Sourceplatform.metix.ai, metix.ai homepage and /pricingread 2026-09-12 |
||
|
APIs, SDKs & MCP Extensibility Composability layer: stable APIs, SDKs, MCP tool consumption/serving, custom tools, and integration into internal systems. |
||
|
HumanlyAPIs, SDKs & MCP Extensibility There is no developer section, API reference, endpoint list, authentication guide, SDK or documentation site, and Humanly names no interface an outside system can call. The two way integrations with nine named applicant tracking and HRIS platforms are prebuilt connectors that Humanly maintains. The Partners program is a go to market program, not a developer one. Sourcehumanly.io homepage, navigation and footerread 2026-09-08 |
||
|
Metix AIAPIs, SDKs & MCP Extensibility A self serve developer platform offers a REST API, an MCP server and four installable agent skills, all under one key. The REST API has the Query Spec query language, natural language search and routes that fetch records by id across people, jobs and companies. A quickstart, a full API reference, an error reference and a changelog support it. The MCP server runs over streamable HTTP at mira-api.metix.ai/mcp on MCP version 2025-03-26, with a legacy SSE endpoint alongside, and Metix says the MCP surface is the REST surface and nothing more, under the same key, limits and credit rules. The four agent skills install with npx skills add MetixAI-Official/metix-skills, work in any agent with or without MCP, and read field names and operators from a live contract route at run time. Rate limits, page sizes and batch ceilings are published, along with the rule that a record takes two calls and the one case where paging can miss a record. Every response arrives in the same envelope, and a query refused for an unknown field is not charged. Access is self serve and ungated. A key is created in about a minute on a free plan with no card and no sales call, and the key status route is free to call. Metix also publishes an llms.txt and a page for AI assistants. Sourceplatform.metix.ai and its documentation indexread 2026-09-12 |
||
|
Testing, Debugging & Optimization Testing, debugging, scoring, retries, fallbacks, quality gates, and optimization loops for improving agent workflows before and after deployment. |
||
|
HumanlyTesting, Debugging & Optimization Every candidate is scored consistently against rubrics and knockout criteria. Humanly names no sandbox, test environment, simulation, versioning, holdout or scored case before deployment, and no way for a customer to try a rubric or knockout change before it runs on live applicants. Humanly positions itself on continuous third party bias auditing of its AI. The audit covers Humanly's own models and is commissioned by Humanly, and Humanly names no way to audit a customer's deployed instance. Sourcehumanly.io/products/ai-interviews and homepageread 2026-09-08 |
||
|
Metix AITesting, Debugging & Optimization A built in quality bar decides which of the agent's candidates reach the customer, and quality is checked at every step of the funnel. Candidates get through only after clearing the customer's bar for fit and real interest. A credit only unlocks a candidate who is qualified and genuinely interested, and if no one qualifies, nothing is charged. Metix gives up revenue whenever its agent's output falls short. Each candidate passes or fails, the result is visible for each candidate, and the customer sets the threshold in the brief. Metix names no test environment or sandbox, no holdout or regression check, no versioning and no evaluator model. It does not say whether a changed search, brief or message template is run against a known set before it reaches live candidates. The published figures are outcome statistics across roles, 95 percent scheduling within 24 hours, 5x the response rate of cold outreach and 65 percent less manual work. Sourcemetix.ai/pricing FAQ and homepageread 2026-09-12 |
||
| Specialist automation | ||
|
Browser & Computer Use Browser, desktop, or remote/local computer control for workflows that cannot be handled through stable APIs alone. |
||
|
HumanlyBrowser & Computer Use The agent works through conversation. Engagement runs over SMS, email, web chat, voice, phone and video, and everything else moves through prebuilt two way connectors into nine named applicant tracking and HRIS platforms plus recruiter and hiring manager calendars. Humanly names no hosted or local browser, desktop session, or remote or local computer control, and no sourcing or capture browser extension. The candidate uses their own browser to take a video interview. Sourcehumanly.io homepage and /products/ai-recruiterread 2026-09-08 |
||
|
Metix AIBrowser & Computer Use Candidates are found and contacted through server side searches and API calls. Search routes run against Metix's own index, records are read by id through detail routes, and outreach and booking run through the customer's approved messages and calendar. Three calls take a role from nothing to records, and nothing in that path opens a page. Metix names no hosted or local browser, desktop session, computer control, RPA or browser extension. Outreach delivery is the one step that touches an outside interface, and Metix names no interface for it. A customer can paste a LinkedIn URL or a job URL when defining a role. The person decides what is read and the product parses it, so no agent navigates a page. Sourceplatform.metix.ai and metix.ai homepageread 2026-09-12 |
||
Pricing snapshot
Sourced from the Index pricing dataset · open each vendor's profile for full detail.
| Pricing | H Humanly |
M Metix AI |
|---|---|---|
|
Entry price Lowest public entry point |
Quote only | $99/month Starter for hiring, with 5,000 credits a month. The developer data platform starts at $0, with paid plans from $9/month. |
|
Pricing confidence How public the numbers are |
Contact only | Public, exact |
|
Billing Primary billing axis |
— | Two separate published systems. The hiring product bills a monthly subscription with an allowance of credits and AI agents, and credits are spent when the agent unlocks a contact (10 credits) or generates a Deep Search list (100 credits). The developer data platform bills monthly credits, where one credit buys 25 search results or 5 records, at rates identical on every plan. |
|
Variable cost Workload / overage exposure |
Medium variable cost | Low variable cost |
|
Free tier / trial Try before you buy |
No free tier
|
Free tierTrial
|
|
Buying motion Self-serve vs sales call |
Sales call | Self-serve |
|
Pricing verified When this pricing was last checked |
More comparisons with Humanly or Metix AI
Other matchups in enterprise operations agents
Not the pairing you were after? These compare a different set of enterprise operations agents on the same 14 capabilities.