Be the recruiter who fills reqs faster — with JDs, InMails, and scorecards built from a real brief, every time.
Prompt Architects stores each req's role context, hiring manager expectations, and tone standard as a Global Variable — so your job description, sourcing InMail, screening questions, and interview scorecard pull from a real intake call brief, not a blank page. You review and approve every draft before anything reaches a candidate or your ATS. So you move every req forward faster without rebuilding the same context from scratch for every new opening.
Free to start · No credit card · No API key · Your prompts stay yours
Textio's analysis of ChatGPT-written recruiting outreach found the output is filled with phrases — "I came across your profile," "cutting-edge" — that lower candidate response rates. The model has no access to data on what passive candidates actually reply to. — Textio — ChatGPT Writes Recruiting Mail
Your JD prompt — enhanced
You type
write a job description for a senior backend engineer
Prompt Architects sends
Role: Senior technical recruiter at [COMPANY_NAME] — [COMPANY_STAGE], hiring ICs for the [TEAM_NAME] team.
Req: [REQ_TITLE] — Senior Backend Engineer, [LEVEL], reporting to [HIRING_MANAGER].
Context: [TECH_STACK]; [SCOPE_OF_WORK]; remote-friendly.
Deliverable: Full job description for careers page and LinkedIn — responsibilities, must-haves, nice-to-haves, and growth path.
Format: Role impact first → responsibilities (6–8 bullets) → required vs. preferred qualifications → compensation and benefits → why us.
Constraint: Inclusive-language pass — no gendered terms, no years-of-degree gatekeeping, no 'ninja' or 'rockstar'; describe the actual work.
[COMPANY_NAME], [TEAM_NAME], [HIRING_MANAGER], [TECH_STACK], and [LEVEL] fill automatically from your saved req profile — no re-briefing the AI on context already captured in your intake call notes.
01 · The problem
Every req starts with a blank ChatGPT window — and the intake call notes are still in a different tab.
50%
Half of Prompt Architects customers had no prompt system before signing up — not a Notion doc, not a shared folder, nothing. Every JD, every sourcing InMail, every interview scorecard started from a blank ChatGPT window with whatever context they could reconstruct from memory or notes.
Prompt Architects customer data, July 2026
Clichés sink response rates
Textio's analysis found that ChatGPT recruiting outreach defaults to phrases like "I came across your profile" and "cutting-edge" that lower candidate response rates — because the model has never been trained on what passive candidates actually reply to. Generic context in, generic clichés out. Candidates receiving dozens of InMails a week recognize and archive them in seconds.
1 default context
ChatGPT's Custom Instructions hold one global context for every conversation. There is no per-req mode. Switch from the backend engineer req to the DevOps opening to the offer letter draft, and you either overwrite the default context, lose the role brief, or paste the intake call notes manually — every session, every tab, every time.
4+ artifacts per req
A single open req typically needs at least four AI-drafted documents: job description, sourcing InMail, screening questions, and interview scorecard — plus follow-up messages and pipeline updates as the req moves through stages. Without saved templates and role variables, each one starts from a blank window. Every new opening is four or more re-briefings minimum.
Prompt Architects customer data, July 2026
You finish the intake call. You have what you need: the seniority level, the tech stack, the two or three things that actually separate a good candidate from the right one. You open ChatGPT. The window is blank. Before you write a word of the JD, you spend ten minutes reconstructing context — who the company is, what the team does, what the role reports to, what the hiring manager said 'strong' looks like. That context was in the intake call. Now it is in a notes tab. You are the integration layer between the two, and you do it for free, for every req.
You source twenty candidates for the backend req. You open ChatGPT to write the InMail. The output reads like every other AI-generated message passive candidates received this week: a company intro that could apply to any team, a hook that is indistinguishable from every other recruiter using the same tool, a close that could have been written for any role. Your candidate pipeline thins. Two weeks later the hiring manager asks why they haven't seen more names. The InMails you sent looked right. They just weren't built from a real req brief, and ghosting followed.
You have eight open reqs. Each has its own hiring manager, its own definition of 'qualified,' its own candidate-facing positioning, its own target for time-to-fill. Your ATS holds the job posting. Your email holds the intake call notes. ChatGPT holds nothing from session to session. You are rebuilding the same role context every single time you open a new prompt window — for every JD, every InMail, every scorecard, every follow-up sequence. None of that rebuild time appears in any metric. It just disappears.
You don't have an outreach volume problem. You have a context storage problem — every req starts blank because no tool is keeping your role knowledge between sessions.
02 · The solution
Now imagine a recruiting toolkit where every req starts fully briefed.
- 1
Save each req's intake call context in under 2 minutes
Enter the role title, hiring manager, seniority level, tech stack, must-have qualifications, and tone standard as a Global Variable. Prompt Architects stores it permanently — available for every JD, InMail, screening question set, and scorecard you draft for this req, across every session and every tab, from now on.
- 2
Pull a template — the req variables fill every placeholder
Choose a job description, sourcing InMail, cold-outreach sequence, screening question set, or interview scorecard from your library. Your req variable fills every placeholder automatically — company context, seniority level, tech stack, candidate-facing tone, hiring manager's must-haves. The AI works from your actual brief, not a generic internet average. Flip the Tone Selector: candidate-facing for warm and persuasive, internal for crisp and criteria-led.
- 3
Review and approve — nothing reaches a candidate or your ATS without you
Enhance the prompt and get a structured first draft with role context, format, and inclusive-language constraints already applied. You read it. You edit it. You send it. Your req variables, saved templates, and tone settings are never used to train any AI model.
Free to start · No credit card · No API key · Your prompts stay yours
03 · What you get
Every req, every stage — one system that already knows the role
Role context saved once — every artifact for that req starts briefed
Global Variables store each req's title, hiring manager, seniority level, tech stack, must-haves, and candidate-facing tone from the intake call. Every JD, InMail, screening question set, and scorecard template pulls them in automatically — no re-pasting notes, no context drift between sessions. The third JD of the morning is as sharp as the first. Switch reqs: the AI already knows the difference.
JD, InMail, screening questions, and scorecard — one library for every stage
A searchable library of job description templates, sourcing InMails, follow-up sequences, cold-outreach scripts, structured screening question sets, and interview scorecards — each built to fill with your req's actual context. Pull the template, the req variable populates it, hit Enhance. Four or more artifacts per req, none of them starting from a blank page.
InMails that read like you wrote them for one candidate — not a list of twenty
The Prompt Enhancer builds outreach from your actual req brief — role-specific context, the right seniority signal, the reason this passive candidate fits — so the output doesn't default to clichés that candidates recognize and archive on sight. Ghosting follows generic outreach. Brief-driven InMails are the difference.
Candidate-facing and internal copy in the right register — without two template sets
Switch from the warm, direct tone that converts passive candidates on LinkedIn to the structured, criteria-led language your hiring managers need to evaluate a finalist scorecard — without rewriting the req context into a separate template. The Tone Selector handles the register shift. You handle the judgment call.
Hiring manager context that travels with every intake call template
Pin the HM's must-haves, team norms, or evaluation criteria from the kickoff email as a Context and it injects into every prompt for that req. No re-pasting pages of hiring manager feedback before you can write a screening question or revise a scorecard. Open the template — the HM context is already there.
You draft. You approve. Your team decides — nothing auto-sends to a candidate
The Chrome extension runs inside ChatGPT, Claude, and Gemini — the tools you already use. No new tab to manage mid-req, no API key, no ATS integration project. Your saved req variables and templates are not used to train any AI model. Prompt Architects drafts. Every JD, InMail, and scorecard is reviewed and approved by you before it goes to a candidate, a hiring manager, or your ATS.
04 · A Monday on you
Three open reqs. Intake call at 2pm. Everything moving before lunch.
It is Monday, 8:52am. Three things are waiting before the calendar fills: the backend engineer JD needs to go to the hiring manager for sign-off by 11am, two passive candidates the sourcer flagged on Friday are waiting for InMails before they accept something else, and the DevOps scorecard the panel used last quarter is the wrong level for the new req — it needs an update before Wednesday's interviews.
On any other Monday, this is three blank ChatGPT windows. The backend JD needs the full intake call context re-pasted before you can write the first bullet. The InMails need the req brief, the candidate-facing tone, and hooks specific enough that a passive candidate with options actually opens them. The scorecard update needs the original, the hiring manager's revised criteria, and the new level benchmarks — three separate sessions, all starting from scratch. Meanwhile the 2pm intake call is for a fourth opening: a Staff Engineer req the CTO flagged as a priority hire.
You have not yet opened your inbox.
Not today.
Not today.
The backend req variable is already saved: role title, seniority level, tech stack, reporting line, and the two non-negotiables the hiring manager stated in the intake call. You open the JD template — the req slot fills in. Hit Enhance. Three minutes later you have a structured first draft with inclusive-language constraints applied, ready to edit and route for sign-off. Done by 9:12.
The InMails for the two passive candidates take eleven minutes. Open the sourcing InMail template with the backend req variable loaded. Add the specific context for each candidate from the sourcer's notes. Enhance — the output references the actual role, the actual stack, the actual reason this person might care about this req, not a generic opener they will archive in two seconds. You adjust the first line for each one in your own words and queue them.
The DevOps scorecard update is done before 10:30am. Pull the scorecard template with the DevOps req variable loaded, paste the hiring manager's revised criteria into the context field, hit Enhance — updated competencies organized by must-have vs. nice-to-have, behavior-based question prompts calibrated to the new level. You send it to the panel with a two-sentence note. You spend the time before the 2pm intake call prepping the Staff Engineer screening questions.
You are the recruiter whose candidate pipeline moves — not because you work more hours, but because your system already knows every req before you open the first tab.
05 · Used by hiring teams and solo recruiters who stopped starting from scratch
Used by hiring teams and solo recruiters who stopped starting from scratch
4.9/5 from 150+ verified reviews — including hiring professionals and team leads who built a system for their prompts instead of starting over every session.
Getting the most from my kick off prompt
“Works great to help me get the most from my kickoff prompt. I type in what I think is a great prompt and this turns it into a fantastic prompt. Saves me at least 5 follow up prompts!”
Calms my prompt chaos!
“One of the best things about this product is how much it calms my prompt chaos. I had prompts EVERYWHERE — Notion pages, Google Docs, membership areas, notepads on my phone, bookmarks. Now I have a single source of truth for my prompts! I love that it comes with a library of prompts too. I no longer fight with my wording, and the extension is the cherry on top. This is slowly becoming my favorite purchase, and that's saying something!”
Architectural fix shipped in one day
“Nafiul surprised me — he said the change would take a week or more, but less than 24 hours later he emailed a video walkthrough of a working implementation called Global Variables. The "build once, reuse everywhere" behavior I described as missing is now real on the web app and in the Chrome extension. The product I described in my original review is not the product I have today. Genuinely impressed.”
This extension is weirdly good.
“A few weeks ago I installed Prompt Architects on a whim — didn't think much of it. Now I use it every single day. Turns out the problem with AI isn't the AI — it's that I explain things like a caveman and expect Shakespeare back. This thing takes my messy half-thoughts and restructures them into actual prompts. One click. Done. No more rewriting the same thing four times. The prompt library is genius — I save structured prompts by category and reuse them. Clean UI, no bloat. Just does the thing.”
Workflow Game Changer for Prompt Writing
“This tool completely changed my workflow. It saves me hours every week because I no longer need to write prompts from scratch regularly. The interface is simple, easy to learn, and very fast to adopt into daily work. I can now create high-quality prompts much faster and more efficiently. The template system is extremely useful for repetitive tasks and content generation. I genuinely love this tool and would love to see even more templates added in the future. Highly recommended for anyone who works with AI regularly.”
Great for Team Productivity
“I bought Prompt Architects to help my team get better results from AI, and it has been very practical from day one. It makes it easy for them to turn rough ideas into clear, structured prompts without overthinking the process. The team is using it, enjoying it, and already seeing better outputs with less back-and-forth. Simple, useful, and genuinely helpful for anyone trying to bring AI into daily team tasks. Highly recommended.”
06 · The comparison
Because every blank window you open for a req is time your candidate pipeline isn't moving.
ChatGPT forgets every req the moment you close the tab. Your ATS's built-in AI generates copy from a blank slate with no knowledge of your level benchmarks, HM expectations, or tone standard. Prompt Architects holds the req brief and fills every artifact from it — across every stage of the pipeline.
| What in-house and agency recruiters working live reqs actually need | ChatGPT alone | Your ATS's built-in AI | Prompt Architects |
|---|---|---|---|
| Per-req context saved and auto-loaded — no re-briefing from intake call notes | |||
| HM expectations and must-haves auto-filled into JD, InMail, and scorecard | |||
| JD, InMail, screening questions, and scorecard templates in one library | |||
| One-click structure: role context, requirements, and inclusive-language constraint | |||
| Candidate-facing vs. internal tone in one click — no separate template set | |||
| Outreach built from the actual req brief — not ChatGPT defaults | |||
| Chrome sidebar inside ChatGPT, Claude, and Gemini — no new platform or login | |||
| Free to start — no credit card, no ATS integration, no IT ticket |
Free to start · No credit card · No API key · Your prompts stay yours
07 · Straight answers
Straight answers for recruiters
Will AI-generated JDs and InMails sound like our team, or will candidates notice they came from a template?
The Prompt Enhancer adds structure — role context, requirements, format, and inclusive-language constraints — to the brief you wrote from your intake call notes. It does not replace your positioning with a generic script; it makes sure the AI receives the actual req context clearly. Your editing pass adds the final voice and any candidate-specific hook. Every draft goes through you before it reaches a candidate, a job board, or your ATS. Candidates see a message built from a real req, not a ChatGPT default.
Does Prompt Architects connect to our ATS — Greenhouse, Lever, or Workday?
No integration is needed. Prompt Architects runs as a Chrome extension sidebar inside ChatGPT, Claude, and Gemini — it sits next to wherever you draft content in the browser. Draft the JD or InMail in your AI window, then paste the output into your ATS posting, your LinkedIn job draft, or your outreach tool. No API key, no integration project, no IT ticket. It works alongside your ATS, not instead of it. Verify current features and pricing on each ATS vendor's site.
We've heard AI can introduce bias into hiring decisions. Is Prompt Architects a risk?
The honest answer: it depends on what the tool does. Prompt Architects is a drafting tool. It helps you write a job description, a sourcing InMail, screening questions, or an interview scorecard. It does not touch candidate profiles, does not screen applications, does not score or rank candidates, and does not make or influence any hiring decision. The legal and research concerns about AI bias in hiring — including the Mobley v. Workday collective action certified in May 2025, which concerns AI tools that automatically screen and filter job applicants — address screening and ranking tools, not drafting tools. Every recommendation about who advances in your candidate pipeline remains with your team. Two honest best practices: review every JD and scorecard output before it goes live, and keep candidate PII — names, demographic details, protected characteristics — out of your prompts entirely. Describe the role and the criteria, not the candidates.
Is it safe to store req details, hiring manager notes, and role context in Global Variables?
Your Global Variables and saved templates are stored on Prompt Architects' servers, separate from ChatGPT's memory and conversation logs. What you type into ChatGPT, Claude, or Gemini is subject to those platforms' own privacy policies — review their data settings for sensitive content. For your Global Variables: store role context, level benchmarks, tone standards, and tech stack details. Keep compensation bands, confidential headcount rationale, and candidate PII out of saved variables. Apply the same judgment you would to any cloud SaaS. Your saved data is never shared with other users and is not used to train any AI model.
What if Prompt Architects drafts a JD or InMail that misrepresents the role?
Nothing in Prompt Architects auto-publishes or auto-sends to candidates, hiring managers, job boards, or your ATS. Every enhanced draft appears in your ChatGPT, Claude, or Gemini window — where you read it, edit it, and decide whether it goes out. The workflow is always: Enhance, review, edit, then you act. You are the final checkpoint before anything enters your pipeline. EEO-compliant language, accurate job requirements, and any required compensation disclosures are your team's responsibility to review — Prompt Architects does not substitute for that review.
How long does setup take if I have eight open reqs and two intake calls tomorrow?
Under 2 minutes per req. Give the req a name, add the role title, the hiring manager, the seniority level, the key requirements from the intake call, and the must-haves that separate qualified from right. Save it. The next time you open any JD, InMail, or scorecard template for that req, the context fills automatically. Most recruiters set up their three or four most active reqs in one sitting and run their first structured JD before the end of the session.
Stop rebuilding every req from a blank window. Stage the brief once — fill every JD, InMail, and scorecard from it.
Save each req's intake call context as a Global Variable. Build a library of JD, sourcing InMail, screening question, and scorecard templates. Get first drafts the AI builds from a real brief — inside the ChatGPT, Claude, or Gemini window you already have open.
Free to start · No credit card · No API key · Your prompts stay yours