Product · 25 tokens per run
Difficult Conversation Script
You have a hard conversation on the calendar this week (fire a report, break up, tell an aging parent they cannot drive, push back on a boss, end a friendship) and you keep replaying it in your head instead of writing it down. Describe the situation in one paragraph, name the relationship and your goal, and get a rehearsal script back: an opening line you can actually say out loud, three anticipated pushbacks with prepared responses (matched to the type of reaction that person is likely to throw at you), a graceful exit line, five to eight tempting things NOT to say, and a short followup note you can send in writing after. A framing_note runs first as a coach-not-advocate check: if your own framing shows you may be the unreasonable party here, it says so gently before handing you the script.
About this tool
What does difficult conversation script do?
You have a hard conversation on the calendar this week (fire a report, break up, tell an aging parent they cannot drive, push back on a boss, end a friendship) and you keep replaying it in your head instead of writing it down. Describe the situation in one paragraph, name the relationship and your goal, and get a rehearsal script back: an opening line you can actually say out loud, three anticipated pushbacks with prepared responses (matched to the type of reaction that person is likely to throw at you), a graceful exit line, five to eight tempting things NOT to say, and a short followup note you can send in writing after. A framing_note runs first as a coach-not-advocate check: if your own framing shows you may be the unreasonable party here, it says so gently before handing you the script.
How does Difficult Conversation Script work?
Describe the hard conversation you have on the calendar this week in one paragraph (what happened, who is involved, what you have already tried, what you are afraid the other person will say), name your relationship to the other person, name the single outcome you want, pick a tone (compassionate, firm, or balanced), and optionally name one topic you do NOT want raised. One structured AI pass returns a full rehearsal artefact you can read on paper or on your phone right before the conversation: a coach-not-advocate framing_note, a word-for-word opening_line in your voice with delivery notes (breath, pause, in-person vs call, sit vs stand), exactly 3 anticipated pushbacks typed by reaction (anger, deflection, bargaining, tears, guilt trip, silence) each with a prepared response in your voice, a graceful exit line that works whether the conversation went well or badly, 5 to 8 what-NOT-to-say entries tailored to your specific situation, and a 2 to 4 sentence followup note template you can send in writing after.
How is this different from asking ChatGPT to help me plan a hard conversation?
A freeform ChatGPT prompt writes a plausible pep talk, often invents accusations or quotes the other person never said, drifts into scorched-earth ultimatums you did not ask for, and takes your framing at face value even when your own words show you may be the unreasonable party. Difficult Conversation Script runs one structured pass with five hard rules: it never fabricates accusations or quotes about the other person beyond what you described, it never writes an ultimatum you did not ask for, it never centres your other grievances (if you brought one conversation, the script is about that one conversation), it honours one_thing_off_limits strictly (never raises it in any field, references it once in what-NOT-to-say), and it runs a coach-not-advocate check FIRST that gently pushes back if your own framing shows you may be the one out of line. You get a word-for-word opening line and 3 typed pushback responses, not a paragraph of advice.
What does the coach-not-advocate framing note do, and can it refuse to write my script?
The framing_note runs BEFORE the script and picks one of three modes. Clear-eyed: your framing checks out and you have real grounds, framing_note says so in one honest sentence and the script proceeds normally. Gentle pushback: your OWN words reveal you may be the unreasonable party (firing a report as retaliation for a complaint they made, ending a friendship because the friend set a limit you resent, calling someone else's normal reaction 'crazy'), framing_note names what it noticed in one to three sentences without lecturing, and the script still generates but is tuned to that reality. Refuse-and-route: the situation trips a safety trigger (intimate partner violence, safety planning to leave an abuser, suicidal ideation, a plan to harm someone, retaliation against a legally protected activity like an HR complaint or a whistleblower disclosure) and the tool refuses to write a rehearsal script that could put you or someone else at more risk. In refuse mode it points you to the specific resource that CAN help (the National Domestic Violence Hotline at 1-800-799-7233, the 988 Suicide and Crisis Lifeline, a licensed therapist, an employment lawyer) instead of drafting the script.
What are the three anticipated pushbacks and how does it pick them?
Every script ships EXACTLY 3 anticipated pushbacks, one per entry, ideally covering three different pushback types drawn from a fixed list of six: anger, deflection, bargaining, tears, guilt_trip, and silence. They are ordered from most likely to least likely for THIS relationship and this situation, so the first one is what the person is most likely to throw at you and the third is the wildcard. Each entry has expected_words voiced in the OTHER person's likely voice (one sentence) and your_prepared_response written in your voice (one to two sentences), tuned to hold the line without escalating and to honour the tone you picked. This is the wedge over a generic pep talk: you walk in with three responses already rehearsed instead of freezing when the person cries or gets defensive.
Will it work for firing a report, breaking up, or telling an aging parent something hard?
Yes on all three, and the relationship field is the lever that tunes the whole script. A line to a direct report is not a line to a romantic partner, and neither is the line you say to your 78-year-old dad. Free text is accepted so you can be specific ('direct report I inherited', 'romantic partner of 4 years', 'aging parent who lives alone', 'co-founder I no longer trust', 'boss whose approval I still want', 'close friend who set a boundary I resent'). The opening_line, the pushback responses, and the followup note template all register to that relationship so the script reads like something you would actually say, not a template lifted off a management blog. If your situation is one this tool cannot help with (an active safety crisis, suicidal ideation, retaliation for a protected activity), the framing_note refuse-and-route mode says so plainly and points to the resource that can.
What is the graceful exit line, and why does it matter?
The graceful_exit_line is ONE sentence you can say to end the conversation cleanly and it is designed to work whether the conversation went well OR badly. If the other person heard you and it landed, it is the exit that preserves the moment. If the conversation collapsed and the other person exploded or shut down, it is the exit that lets you leave with your self-respect intact instead of scorched-earth. The tool will not draft a scorched-earth exit under any tone, even 'firm', because the exit line is the last thing said and it sets the tone for the followup note template you send in writing after. One sentence in, one sentence out, no hostage-taking at the door.
How much does it cost per rehearsal script, and do I need a subscription?
No subscription. Each rehearsal script costs 25 produde tokens, a prepaid platform credit you buy once and spend across any produde tool, so you pay per hard conversation instead of a monthly therapy-lite or coaching-app fee. If a run fails, the full token cost is refunded to your wallet automatically. Most people have one to three genuinely hard conversations a year (the kind you keep replaying in your head for a week beforehand), so a small token top-up covers a long stretch and there is no minimum spend and no per-script cap.
Who is it for?
Anyone with a hard conversation on the calendar this week that they keep replaying in their head instead of writing down. Managers who need to fire a report and want the opening line rehearsed before they walk in. Adult children telling an aging parent they cannot drive anymore. People planning to end a romantic relationship who want the line ready and the followup note drafted before they lose their nerve. Co-founders having the ownership talk. Direct reports pushing back on a boss whose approval they still want. Anyone ending a friendship, saying something they have avoided saying, or handing off a hard decision. It is a focused single-purpose rehearsal tool for one specific conversation on your calendar; it is not therapy, not legal counsel, and not a safety-planning service (the tool refuses and routes you elsewhere if the situation trips a safety trigger).
What does it cost?
25 tokens per run. Produde tokens are a prepaid platform credit; buy them once and spend on any tool.
Do I pay if a run doesn't complete?
No. You only pay for results. If a run can't finish, its full token cost stays in your wallet automatically.
Is my data stored?
Your input and the run output are saved to your account so you can browse history. They are not sold, shared with advertisers, or used to train shared models. See privacy for details.