interviewflowai-mcp

# InterviewFlowAI MCP

Connect InterviewFlowAI to Claude, ChatGPT, Codex, and other MCP-compatible AI assistants. Ask about your candidates and interviews in plain language, and perform supported recruiting actions without leaving your AI workflow.

```text
"Who are my top candidates for the Senior PM role, and where did each of them shine?"

"Compare our top two finalists on communication and problem-solving,
 and draft a summary for the hiring manager."

"Show me everyone who completed the interview this week,
 grouped into strong, maybe, and no."
```

Ask a question like that in Claude or Codex, and the answer comes back from your own InterviewFlowAI workspace — not from a spreadsheet you pasted in ten minutes ago. The assistant finds the candidates, reads the interviews behind them, and answers the question you actually had.

> **InterviewFlowAI MCP is a hosted remote MCP server.** You do not need to run anything locally, install a package, or manage a server. You connect your AI assistant to `https://api.interviewflowai.com/mcp` and sign in with your InterviewFlowAI account.

- **MCP endpoint** — `https://api.interviewflowai.com/mcp`
- **Registry name** — `com.interviewflowai/mcp`
- **Documentation** — https://docs.interviewflowai.com/platform/mcp
- **Product page** — https://interviewflowai.com/features/mcp

---

## Why this exists

Recruiting teams already use AI assistants to write outreach, summarize notes, and prep for debriefs. The gap is data: the assistant does not know who applied, who finished an interview, or how anyone scored. So you copy and paste, and the assistant reasons about a stale fragment of your pipeline.

The [Model Context Protocol](https://modelcontextprotocol.io) (MCP) is an open standard that lets AI assistants connect to external systems through a consistent interface. InterviewFlowAI MCP is an MCP server for recruiting: it gives your assistant a supported, permission-aware path to the candidate and interview data already in your InterviewFlowAI workspace.

The result is an AI agent for recruiting that answers from live data — a recruiting MCP you can point Claude, ChatGPT, or Codex at.

---

## What can you do with InterviewFlowAI MCP?

With read access (`mcp:read`), a connected assistant can:

- **Find candidates** in your workspace, including by interview completion status.
- **Inspect candidate information** — the supported fields on a Candidate record.
- **Retrieve AI Interviewer information** — list the AI Interviewers configured in your workspace and inspect their details.
- **Review interview information**, including interview scores.
- **Summarize candidate and interview context** so you can read the substance of an interview in seconds instead of scanning a full transcript.

With write access (`mcp:write`, workspace Owners only), a connected assistant can:

- **Update supported candidate fields** — notes, custom fields, visibility, and archived status.

That is the supported surface. InterviewFlowAI MCP does not screen, rank, reject, or advance anyone on its own — it finds, surfaces, and organizes information so a recruiter can decide.

> **A note on tool names.** This repository documents capabilities rather than individual tool names, because the exact tool list is defined by the hosted server and can change between releases. Your assistant discovers the current tools automatically when it connects. See the [official documentation](https://docs.interviewflowai.com/platform/mcp) for the authoritative capability list.

---

## Example prompts

### Find completed interviews

```text
Show me the candidates who completed the interview.
```

### Find top candidates

```text
Show me the top 10 candidates by interview score.
```

### Filter by score

```text
Show me candidates with an interview score of 70 or above.
```

### Going further

Those three are deliberately simple — they are the fastest way to confirm the connection works. Once it does, you can ask for the retrieval and the thinking in one question:

```text
Of the candidates who scored above 70, which three would you look at first,
and what would you want to probe in a live call?
```

```text
This candidate scored lower than the others but I liked them on the call.
What does the interview actually show, and where does the score come from?
```

```text
Across the candidates for this role, what are the most common weak spots?
Is that the candidates, or is it how the interview is asking the question?
```

A larger, categorized set lives in **[examples/recruiting-prompts.md](examples/recruiting-prompts.md)**, and end-to-end recruiting workflows in **[examples/workflows.md](examples/workflows.md)**.

---

## How it works

```text
Recruiter
   │
   ▼
Claude / ChatGPT / Codex / MCP client
   │
   ▼
InterviewFlowAI MCP
   │
   ▼
InterviewFlowAI
   │
   ▼
Candidates + Interviews
```

In plain terms:

1. **InterviewFlowAI hosts the remote MCP server.** It runs at `https://api.interviewflowai.com/mcp`. Nothing to install or operate.
2. **Your AI assistant connects to it using MCP.** Any MCP-compatible client can speak to it over the Streamable HTTP transport.
3. **You sign in with your InterviewFlowAI account.** That sign-in is what tells the server which workspace you are in and what you are allowed to do.
4. **The assistant calls supported InterviewFlowAI tools** on your behalf, and answers your question using what comes back.

Your assistant never gets blanket access to InterviewFlowAI. It gets exactly the access your own account has, and only through the supported tools the server exposes.

---

## MCP endpoint

```text
https://api.interviewflowai.com/mcp
```

| | |
|---|---|
| Transport | Streamable HTTP |
| Authentication | OAuth 2.1, using your InterviewFlowAI account |
| Scopes | `mcp:read`, `mcp:write` |
| Hosting | Hosted by InterviewFlowAI — no local install |

The server implements standard MCP OAuth discovery, so most clients need nothing beyond the URL: they find the authorization server themselves and walk you through sign-in in your browser.

---

## MCP Registry

InterviewFlowAI MCP is published in the official [Model Context Protocol Registry](https://registry.modelcontextprotocol.io/), the standard index that MCP clients and directories use to discover servers.

```text
com.interviewflowai/mcp
```

| | |
|---|---|
| Registry name | `com.interviewflowai/mcp` |
| Version | `1.0.0` |
| Transport | Streamable HTTP |
| Remote endpoint | `https://api.interviewflowai.com/mcp` |

The `com.interviewflowai` namespace is domain-verified, so the listing is published by InterviewFlowAI itself rather than by a third party.

You can query the live record directly:

```bash
curl "https://registry.modelcontextprotocol.io/v0.1/servers?search=com.interviewflowai/mcp"
```

The manifest published to the registry is [`server.json`](server.json) in this repository. Being listed in the registry does not change how you connect — clients that read the registry can find InterviewFlowAI automatically, and everything in [Connect InterviewFlowAI to your AI assistant](#connect-interviewflowai-to-your-ai-assistant) works exactly the same either way.

---

## Connect InterviewFlowAI to your AI assistant

Pick your client below. In every case the only value you need is the endpoint URL, and authentication happens in your browser.

> Setup steps for MCP clients change frequently. The commands below follow each vendor's current documentation, but if a client has changed its syntax, that vendor's own MCP documentation is the source of truth.

### Claude Code

```bash
claude mcp add --transport http interviewflowai https://api.interviewflowai.com/mcp
```

Then authenticate:

```text
/mcp
```

Select **interviewflowai** and complete the browser sign-in. To make the server available across all your projects rather than just the current one, add `--scope user`.

Reference: [Claude Code MCP documentation](https://code.claude.com/docs/en/mcp)

### Claude Desktop

1. Open **Settings → Connectors**.
2. Choose **Add custom connector**.
3. Enter a name (for example, `InterviewFlowAI`) and the URL `https://api.interviewflowai.com/mcp`.
4. Save, then click **Connect** and sign in with your InterviewFlowAI account.

### Codex CLI

Add the server to `~/.codex/config.toml`:

```toml
[mcp_servers.interviewflowai]
url = "https://api.interviewflowai.com/mcp"
```

Then authenticate:

```bash
codex mcp login interviewflowai
```

Reference: [Codex MCP documentation](https://developers.openai.com/codex/mcp)

### Codex IDE extension and desktop app

Open **Settings → MCP servers → Add server**, choose **Streamable HTTP**, enter `https://api.interviewflowai.com/mcp`, and complete the sign-in. Codex IDE and desktop read the same `~/.codex/config.toml`, so a server added through the CLI shows up there too. Restart the application after adding it.

### ChatGPT

InterviewFlowAI can be added to ChatGPT as a custom connector using the same endpoint URL. Connector availability depends on your ChatGPT plan and workspace settings — see the [InterviewFlowAI MCP documentation](https://docs.interviewflowai.com/platform/mcp) for current setup steps.

### Other MCP-compatible clients

Any client that supports remote MCP servers over Streamable HTTP with OAuth can connect. Point it at:

```text
https://api.interviewflowai.com/mcp
```

Most clients will discover the authorization server automatically and prompt you to sign in.

---

## Authentication

You authenticate with **your own InterviewFlowAI account**, using the same email address you sign in to InterviewFlowAI with. Authentication uses OAuth 2.1 through InterviewFlowAI's identity provider, and the sign-in happens in your browser — you do not paste an API key into your AI client, and your password is never handled by the client.

Your session determines two things:

- **Which workspace** the assistant can see. You get your company's workspace, and nothing else.
- **What the assistant may do** in it, based on your role.