0saas

Job ticket 779461 · smb · harbor ledger · no subscription

Nine hundred credits a year. Judgment was never on the meter.

Apollo's Free plan lasts 900 credits a year — 2 sequences, 5,000 AI words a month — and unlimited sequences start at US$49 per seat a month billed annually, with a second counter running in credits. This job ticket hands the qualification pass to prompts you own: paste your ICP and researched account notes, and the ledger berths each account SAIL, HOLD or SKIP with the decisive quotes; run the opener desk and each SAIL account gets an under-90-word first touch whose every claim cites your notes. The database and the sending rails stay Apollo's.

Apollo.io is a trademark of its owner. This is an independent, unaffiliated comparison built on Apollo.io's own published pricing material, linked below and re-read every thirty days.

Paste the notes, press stamp the manifest, and the berths fill — SAIL accounts with one checkable hook each, HOLD accounts with the missing match named, and nothing stamped on evidence you did not paste. — rendered in this browser

Source: noneRendered: 0Uploaded to 0saas: nothing

Sample data is illustrative, not a claim about Apollo.io. Every Apollo.io figure on this page comes from the linked pricing page, read 2026-09-27.

The prompt pack

Two prompts, each mapped one-to-one onto a Apollo.io feature you are currently rationed on. Run them from this page with a free key, or copy them into any free chat model — the text is identical either way, and it is printed in full at the bottom of this section so it survives this site.

Loading the pack from tools.json… If this fails, the full prompt text is still readable below.

One-time unlocks only, never a subscription. If a price ever appears here it was read from Stripe in that request, not typed by hand. An unlock is client-side gating, not secure payment verification: it is stored in this browser and cannot prove payment to anyone else.

What Apollo.io's own pricing page says

Read from apollo.io/pricing on 2026-09-27. Fields we could not read are marked unconfirmed rather than filled in with a plausible number.

What Apollo meters — read from apollo.io/pricing on 2026-09-27 — vs this job ticket. — verified 2026-09-27
The jobApollo FreeThis pageThis page + a free-tier model
Annual credits / sequences900 / 2Not a product — judgment over your own notesundefined
Account qualificationCredit-gated database lookupsFit stamps from pasted notes with quotes, US$0undefined
AI-written openers5,000 words/month on FreeGrounded, cited openers per SAIL accountundefined
Contact database · sendingThe paid platformNot displaced — stays Apollo'sundefined

What this replaces, and what it does not

A displacement claim you cannot substantiate is a lie with good typography. This is the whole scope, including the parts where Apollo.io wins.

A stamped manifest

SAIL / HOLD / SKIP per account with dual quotes and named hooks

Cited openers

Under 90 words, observation-first, every claim resting on a quoted note

A you-own export

CSV and Markdown downloads of the whole manifest

No contact database

You bring notes from your own research; nothing is enriched

No email sending

Openers download as text; sending happens on your rails

No deliverability tooling

Warm-up, tracking and sequences-as-infrastructure stay with platforms

The prompts, in full, as text

Copy these into any model. They are the product; the page around them is convenience. Each block is the exact system prompt the interactive runner sends, plus the user-message template.

Prompt 01 — 1. Stamp the manifest · apollo-fit-stamp v1.0.0 · best ally: Groq Cloud — free tier · llama-3.3-70b-versatile

Replaces: The qualification layer before credits are spent: your pasted account notes berthed on quays — SAIL (sequence-worthy), HOLD (not yet) or SKIP — each with quoted reasons and its entry hook.

Why this model: Fit qualification is binary-plus-reasons pattern matching against your stated ICP; a free 70B at low temperature stamps conservatively, only from pasted notes, and never inflates a marginal account. Fallback ally: OpenRouter — :free model variants.

System prompt (1406 chars):

You keep the harbor ledger: the reader pastes their ideal-customer profile (two or three lines) and their account notes — one paragraph per prospect from their own research. You berth each account: SAIL, HOLD or SKIP, the stamp's reasons quoted from the note, and for SAIL accounts the entry hook — the one concrete thing worth opening with.

Return ONLY one JSON object. No markdown fences, no commentary, no preamble.

{
  "kicker": string (max 8 words),
  "title": string (max 8 words),
  "subhead": string (max 20 words),
  "items": [
    { "title": string (the account name as pasted; body: the stamp's reasons in <= 15 words with the decisive note quoted, then the entry hook (SAIL only: one concrete, checkable fact from the note)), "tag": string (one of SAIL | HOLD | SKIP), "body": string }
  ],
  "notes": [ string (one short footnote per run-wide decision) ]
}

Hard rules:
1. Every stamp names which ICP line it matched or failed — quotes from both the ICP and the note.
2. SAIL requires at least two ICP matches evidenced in the note; one match is HOLD with the missing match named.
3. Nothing is researched for you: absence of evidence in the paste caps the stamp at HOLD.
4. Hooks are facts (an open role, a launch, a quoted goal), never flattery templates.
5. Duplicate accounts merge with the fuller note surviving; the merge is stated.
6. Write stamps in the language of the pasted notes.

User message template:

Stamp my manifest. My ideal customer profile (2-3 lines): {{icp}} My account notes (one paragraph per prospect): {{accounts}}

Variables:

  • {{icp}} — Your ICP (max 300 chars) sample data shipped with the tool
  • {{accounts}} — Account notes (paste) (max 4000 chars) sample data shipped with the tool

Output schema (the answer must match this exactly):

{ "type": "object", "required": [ "items" ], "properties": { "kicker": { "type": "string" }, "title": { "type": "string" }, "subhead": { "type": "string" }, "items": { "type": "array", "minItems": 1, "maxItems": 60 }, "notes": { "type": "array" } } }

Decoding: temperature 0.2 · max 2000 output tokens · responseMimeType application/json · budget ~1,200 in / ~1,000 out for ten accounts · cost per run: 1 request against Groq's free daily quota; the relay floor needs no key at all.

Prompt 02 — 2. Draft the grounded openers · apollo-opener-desk v1.0.0 · best ally: OpenRouter — :free model variants · qwen/qwen3-32b

Replaces: The 5,000-word ration, freed: one first-touch email per SAIL account — under 90 words, one observation from your notes, one question — no invented familiarity.

Why this model: Grounded outreach is evidence-citation dressed as conversation; a free 32B writes tight per-account openers when every sentence must cite the pasted note, which is precisely the constraint that kills template smell. Fallback ally: Groq Cloud — free tier.

System prompt (1425 chars):

You run the harbor's opener desk: given the reader's SAIL accounts (notes + hooks) and what they sell in one line, you draft one first message per account. Each opener is under 90 words, leads with the hook as an observation, states the seller's relevance in one sentence, and ends with one small question. Every fact cites the note; nothing pretends to a relationship that does not exist.

Return ONLY one JSON object. No markdown fences, no commentary, no preamble.

{
  "kicker": string (max 8 words),
  "title": string (max 8 words),
  "subhead": string (max 20 words),
  "items": [
    { "title": string (OPENER: the account name; body: the full message, then the quoted note-fact each of its claims rests on. WATCH: one line naming the.template risk this draft specifically avoids), "tag": string (one of OPENER | WATCH), "body": string }
  ],
  "notes": [ string (one short footnote per run-wide decision) ]
}

Hard rules:
1. Observation first, pitch second, question last — the order is structural.
2. No template tells: 'I hope this finds you well', 'I came across', 'quick question' are banned openers.
3. The question must be answerable in one sentence and about their world, not your meeting.
4. Claims cite note-facts quoted at the end of the opener; an uncited claim deletes itself.
5. Under 90 words per message including greeting and sign-off.
6. Write in the reader's pasted language and brand tone if given.

User message template:

Draft openers for my SAIL accounts. What I sell (one line): {{sell}} The accounts with their notes and hooks: {{sail}}

Variables:

  • {{sell}} — What you sell (one line) (max 160 chars) sample data shipped with the tool
  • {{sail}} — SAIL accounts + hooks (paste) (max 3000 chars) sample data shipped with the tool

Output schema (the answer must match this exactly):

{ "type": "object", "required": [ "items" ], "properties": { "kicker": { "type": "string" }, "title": { "type": "string" }, "subhead": { "type": "string" }, "items": { "type": "array", "minItems": 1, "maxItems": 60 }, "notes": { "type": "array" } } }

Decoding: temperature 0.5 · max 1800 output tokens · responseMimeType application/json · budget ~1,000 in / ~1,000 out for five openers · cost per run: 1 request against the free OpenRouter daily pool; the relay floor needs no key at all.

Questions worth asking before you cancel

Isn't Apollo's value the database?

The database is its product and this page leaves it untouched. What the meter also rations — deciding who deserves outreach and writing the opener — is judgment over notes you already own, and that part never needed 900 credits.

Why must every opener claim be cited?

Because uncited claims are where template smell breeds. When each sentence must name the note-fact it rests on, flattery templates fail the rule and die — what survives is observable, checkable and short.

What does 'monthly figures did not render' mean for the prices?

Apollo's pricing page showed the annual-billing view ('Save 24%') when read on 2026-09-27; the month-by-month figures were not in the capture, so the artifact marks them unconfirmed rather than guessing them.

Where do my prospect notes go?

Into one request to your free ally and nowhere else — no CRM, no enrichment calls, no stored records. The manifest downloads to you; the harbor forgets the fleet when the tab closes.

Nothing is loaded from X or LinkedIn until you click, and no share is counted, logged or reported back to us — the buttons are plain links to each network's own share page.

Next on the ledger