Which subscriptions you can actually replace with something free

You are helping me decide whether to replace one paid subscription with a free alternative. Your bias should be against switching, because a bad swap costs me more than the subscription did, in time and sometimes in lost data. Tell me to keep paying when that is the right answer, and say it plainly.

FILL THIS IN FIRST
- The subscription: [name, and the exact plan I am on]
- What I actually pay: [amount, currency, and monthly or annual]
- How long I have had it: [roughly]
- The three things I genuinely use it for: [be specific. Not "design", but "resizing the same social post into four formats"]
- The last time I used each of those three: [this week / this month / cannot remember]
- What lives inside it: [files, documents, a library, history, nothing]
- How much I would mind losing that: [a lot / a bit / not at all]
- How technical am I willing to get: [none, I want an app I open / some, I can follow setup instructions / a lot, I can run a server]

ASK ME FOR ANYTHING BLANK BEFORE YOU ANSWER, especially the three real uses and when I last used them. Most subscriptions survive an audit because the owner described the product rather than their own use of it.

THEN DO THESE FOUR JOBS IN ORDER

JOB A: WHAT AM I ACTUALLY BUYING.
Take my three real uses and say, for each, whether the value is in software that could be rebuilt, or in something that cannot: a library of content someone was paid to make, a network of other people, a licence, a catalogue, a device integration, or a support guarantee. Be explicit about which of my uses fall on the cannot side, because those are the ones that end the conversation.

JOB B: THE CANDIDATES.
Name up to three free or cheaper alternatives. For each one, give me:
- What it does not do that my current one does, in plain terms, tied to my three uses.
- Whether it is genuinely free or a trial. If you can search the web, open the vendor's own pricing page and give me the figure, the URL and the date you read it. If you cannot search, write "[CHECK THE PRICING PAGE]" next to any price or limit and name the page for me to open.
- How technical it is, against the level I gave you.
- Whether it is run by a company, a funded project, or one person, and say if you do not know.
Do not list something you are unsure still exists. If you are not confident it is current, say so and tell me to check.

JOB C: THE EXIT TEST.
- Can I export my data out of the current one, in what format, and does the alternative import that format? If either half is no, skip Job D and give the KEEP PAYING verdict now, saying it was the export that decided it.
- What would I lose in the export: history, versions, comments, formatting, links?
- If the free alternative shut down tomorrow, could I get my data out of that one too?

JOB D: THE SHELF LIFE.
Tell me what to check before I trust the free tier: when this alternative last changed its pricing, whether the free tier has ever been reduced, who pays for it, and if it is open source, when the project last had activity. Give me the specific places to look, not general advice. Say plainly that free tiers are withdrawn regularly and that this is the check most people skip.

FINISH WITH A VERDICT IN THIS SHAPE
- REPLACE, TRY IN PARALLEL, or KEEP PAYING. One of those three, no hedging.
- One sentence of why, referring to my three uses.
- If REPLACE or TRY, the first three steps, in order, starting with exporting my data.
- The one thing most likely to make me regret this, stated plainly.

NEVER
- Never quote a price, a storage limit, a free-tier cap or a feature from memory. Either read it off the vendor's page now and give me the URL and the date, or mark it for me to check and name the page.
- Never recommend a tool you are not confident exists today.
- Never tell me to cancel before the replacement has been tested on real work.

I have decided to try replacing a paid subscription with a free alternative. Build me a plan that proves it works before I cancel anything. You are not here to cheer me on. You are here to define, in advance, what would make me abandon this, so that I notice it when it happens instead of talking myself round.

FILL THIS IN FIRST
- Leaving: [current tool and plan]
- Trying: [the free alternative]
- Billing: [monthly or annual]
- Renewal date: [date, or "not sure, I will check"]
- The three real jobs it has to do: [the same three I use the paid one for]
- What would count as failing: [my words, if I have any]

IF I HAVE NOT GIVEN YOU A RENEWAL DATE, TELL ME TO FIND IT FIRST. The whole plan hangs off it.

BUILD ME THE PLAN IN FOUR PARTS

PART 1: BEFORE ANYTHING ELSE
- Export everything from the current tool, in the most portable format available, and store the export somewhere that is not either tool.
- Confirm the export actually opens. Tell me to open it, not just download it.
- Write down what the current tool costs and when it renews.

PART 2: THE TEST, WITH A PASS MARK
For each of my three real jobs, write a specific test: the actual task, done on real work, not a demo. For each one, define what PASS looks like before I start, in a way I could not argue my way around afterwards. Include how long it took me on the old tool, so the comparison is honest about time as well as money.

PART 3: THE CHECK-INS
If I am on a monthly plan, four check-ins, one a week, running up to the renewal. If I am on an ANNUAL plan, do not tell me to wait a year: start the parallel run four weeks before the renewal date and use the same four check-ins, and say plainly that the overlap costs me time rather than a second payment. Each check-in asks the same three questions: what did I use the new one for, what did I go back to the old one for, and what took longer than it should have. Tell me that going back to the old one even once is data, not failure.

PART 4: THE DECISION, ON A DATE
Set the decision for a specific number of days before the renewal date, with enough margin to cancel properly. Give me the three outcomes:
- All three jobs passed and I did not go back: cancel, and set a reminder to check the export still opens in six months. On an annual plan, make sure that decision lands before the renewal charge rather than after it.
- One job failed: name the specific question I have to answer, which is whether that job is worth the whole subscription.
- Two or more failed, or I kept going back: keep paying, and say so without treating it as a waste. Knowing was cheap.

ALSO TELL ME
- The cancellation route for the old tool, and warn me that some are deliberately awkward, so I should start it a few days early.
- Whether cancelling loses my data immediately or at the end of the term.
- Any hidden cost in the free option: hardware, my own time to maintain it, or a paid add-on for the part I actually need, such as syncing between devices.

NEVER
- Never assume a price, a renewal window or a cancellation policy. Read it from the vendor's page and cite it, or mark it "[CHECK THIS YOURSELF]" and tell me where.
- Never tell me to cancel early to save a month. The overlap is the point of this plan.
- Never let me count the saving before the test has passed.