AI that drafts, screens and answers, while people decide
Anuva puts AI where hiring takes the most time: writing roles, reading CVs, drafting replies and pulling numbers. Every score shows its evidence and nothing is sent without a person.
What the AI does today
AI CV screening
Each CV is read against the role's scoring template, with the evidence shown for every criterion. The template weights and cutoff set the score.
Job descriptions from one line
Describe the role in a sentence and AI drafts the whole description. When you save it, AI drafts a scoring template from the final text.
Past applicants who fit
Open a new role and Anuva surfaces people from earlier roles who match it, so strong runners up get a second look.
An AI assessment on each application
A written assessment sits on every application next to the score, for the reviewer to read before deciding.
Drafts for emails and invites
The assistant drafts candidate emails and meeting invites. Drafts never send or act on their own. A person approves them.
Ask your reports in plain English
Ask things like how many offers went out last month. Answers come from a fixed set of safe questions, never free form database access.
Guardrails that come switched on
Off by default
AI screening stays off until an admin turns it on for a role.
Your key or ours
Use the platform model, or bring your own OpenAI compatible key.
A person decides
The template sets the score and a person makes every hiring decision.
Fairness report
Selection rates by group, checked against the four-fifths rule.
Built to work with your own agents
REST API with workspace keys
Create and revoke API keys in settings, then read and write your hiring data from your own tools.
Webhooks
Subscribe to events such as application.hired, which carries the candidate package, and start your own workflows.
Careers sites AI can read
Every careers site publishes llms.txt and Google for Jobs structured data, so AI assistants and search engines can read your open roles.
MCP server for Claude, ChatGPT and Cursor
Let an AI assistant list roles, read applicants, summarise the pipeline, add notes and schedule interviews, within the permissions you choose. Set it up
Illustration of an assistant using the Anuva MCP server. Names are examples.
AI questions
Something else? Write to hello@anuva.io and a person will answer.
Does AI decide who gets hired?
No. AI reports evidence and drafts text. The role's scoring template sets the score, and a person makes every shortlist, reject and hiring decision.
Which AI models does Anuva use?
AI screening works with OpenAI compatible models. Use the platform model within your plan's monthly AI screens (Free gets one-time trial credits), or bring your own provider key. Anuva's key routes only to providers that keep no copy of your data.
Can the assistant send emails by itself?
No. The assistant drafts candidate emails and meeting invites. Drafts never send or act on their own, and a person approves each one.
Can our own agents work with Anuva?
Yes. Create workspace API keys in settings to use the REST API, and subscribe to webhooks such as application.hired, which carries the candidate package.
Is there an MCP server?
Yes. Anuva runs a remote MCP server at app.anuva.io/mcp. Connect Claude, ChatGPT or Cursor by signing in to Anuva, or use a workspace API key. The assistant can only do what the permission set you choose allows.
Can AI assistants read our open roles?
Yes. Every careers site publishes llms.txt and Google for Jobs structured data, so AI assistants and search engines can read your open roles.
Try the AI on a real role
Start free, publish one role and turn on AI screening when you are ready. Connect your AI assistant through the MCP server in a few clicks.
or email hello@anuva.io



