# Bot2Bot Labs — extended llms.txt
Canonical site: https://bot2botlabs.com
X: https://x.com/bot2botlabs

Dated AI receipts: what AI-built software costs to build and run, how long it took, what broke, and what subscription it replaces.

## Receipt Scan v1.0 (published 2026-10-30)
Bot2Bot Labs measures AI receipts in five dimensions. Full tables: https://bot2botlabs.com/methods. JSON: https://bot2botlabs.com/methods/receipt-scan-v1.json.

1. Cost to build — builder-stated hours/cost with source+date, or our timed rebuild; estimated vs billed tokens separate (measured_by_bot2bot_labs / builder-stated).
2. Cost to run — smallest working tier; monthly cost at 1, 10, 100 users from cited pricing (price_date).
3. Reliability — fixed core-flow tests (passed/total), concurrency burst, setup-from-README, 14-day breakage log.
4. Security vs real product — dependency audit counts, secrets/auth/headers/passive scan on OUR copy only; vendor compared via public security/trust/cert pages (SOC 2, ISO 27001) cited — we never scan vendor systems.
5. Feature parity — checklist vs vendor public feature list (matched/total).

Field labels: builder-stated | measured_by_bot2bot_labs | unverified. Overall scan: unverified | partially_verified | verified (legacy receipt field verification: measured | partially_verified | reader_reported).

Optional schema block receipt_scan on receipt.v1.json. Limits: without builder/subscriber access we cannot know private infra, incidents, or paywalled features.

## Key pages
- Home: https://bot2botlabs.com/
- Methods: https://bot2botlabs.com/methods
- Receipts index: https://bot2botlabs.com/receipts
- Clones index: https://bot2botlabs.com/clones
- Price table: https://bot2botlabs.com/prices
- Calculator: https://bot2botlabs.com/calculator
- Submit a receipt: https://bot2botlabs.com/submit
- Newsletter: https://bot2botlabs.com/newsletter
- Data & schema: https://bot2botlabs.com/data
- About: https://bot2botlabs.com/about
- Disclosures: https://bot2botlabs.com/disclosures
- Privacy: https://bot2botlabs.com/privacy

## Planned receipts (not published — estimates only)
- #1 Landing page: est. tokens $1–10, est. hours 1–3. Cursor agent, one model, Next.js + Resend signup, deploy to Vercel.
- #2 Contact-form bug fix: est. tokens $0.50–5, est. hours 0.5–1.5. Seed a known bug; agent gets only the bug report; log attempts until tests pass.
- #3 Add auth: est. tokens $3–20, est. hours 2–5. Email/OAuth login with a standard library; security check on sessions and secrets.
- #4 CSV-to-dashboard: est. tokens $2–15, est. hours 1–4. Public sample CSV; filterable chart dashboard; time to first correct chart.
- #5 ToS-respecting scraper: est. tokens $2–15, est. hours 2–4. Public API or ToS-allowed target; rate limits and robots check documented.
- #6 Unit tests for an existing repo: est. tokens $3–25, est. hours 1–3. Fixed commit on a small OSS repo; coverage before/after.
- #7 Same feature on 3 models (Cursor vs Claude Code): est. tokens $15–80, est. hours 4–8. One spec, one commit; variance report with full reproducible setup.
- #8 Security review of a vibe-coded app: est. tokens $2–20, est. hours 3–6. Our own app only; checklist incl. prompt-injection into tools.
- #9 Calendly-style booker vs subscription: est. tokens $10–50, est. hours 6–12. 2–4 week runtime log; compare TCO to dated scheduling SaaS price.
- #10 Jordan's real week (day-map): est. tokens $0–10, est. hours 2 (write-up). One week export; AI-assisted vs manual hours; redacted.

## Published receipts (real)
- r-0001: SnagTime, the AI-built Calendly clone: what it really costs to run (partially_verified, price_date 2026-10-17). Human hours: Not disclosed. Est. build tokens $14,982.96. Headline: $28.80/mo run vs $10.00/seat/mo. Key facts: On price date 2026-10-17, estimated monthly run cost is $28.80 (estimate (cited list prices; not a live deployment bill)) versus $10.00/seat/mo for the subscription it replaces. (2026-10-17 https://calendly.com/pricing) Self-hosting costs less than Calendly Standard only at 3+ seats ($28.80 / $10 = 2.88 seats). For one person it costs about $18.80/mo more, before any maintenance time. (2026-10-17) Local setup on a fresh Linux box: About 2 minutes wall-clock including one fix (npm ci: 24 s; 547 packages). (2026-10-17 https://github.com/nateherkai/snagtime/tree/1c95490a4bfb498084fcb8295439befe194c229a) Unit tests: 196/196 pass on Node 24 (one flaky failure on the first run); 2 of 30 test files cannot load on Node 20.19.2. (2026-10-17 https://github.com/nateherkai/snagtime/tree/1c95490a4bfb498084fcb8295439befe194c229a) Build cost on price date 2026-10-17: Not disclosed human hours, $14,982.96 estimated token dollars, USD 350 billed. (2026-10-17 https://x.com/nateherk/status/2093335489231102341) Operator Y1 @1 user: $695.60 (measured_by_bot2bot_labs). User clone: free_self_host; real product @1 seat: $10.00/mo (price_date 2026-10-17). Break-even: 3+ seats — $28.80/mo clone infra ÷ $10/seat/mo Calendly Standard = 2.88 seats → round up to 3+ seats before monthly run cost beats list subscription (= headline.breakeven). Verdict good_for: Teams with 3+ seats who can self-host, patch dependencies, and accept you as support. | skip_if: Single seat, need vendor SOC 2 / SLA, or cannot upgrade Next.js before a public deploy. | main_risk: Critical/high npm advisories on pinned commit; no repo commits since 2026-08-28 (reliability_notes + maintenance_pulse). Confidence: 40% measured, 50% builder-stated, 10% unverified (4/10 key fields). Maintenance pulse (2026-10-17): last commit 2026-08-28T12:30:56Z, issues 3, stars 388, license MIT. Data: Self-hosted PostgreSQL in production Compose (setup.environment); local demo uses SQLite on your machine — not Calendly cloud. Export: partial. Price history (3 rows, re-check every 90d): Calendly Standard (per seat, billed yearly) $10 on 2026-10-17; DigitalOcean Basic Droplet 4 GiB / 2 vCPU $24 on 2026-10-17; Resend Pro $20 on 2026-10-17. Labor-adjusted operator TCO Y1 @1: $695.6 (Same as cash TCO $695.60: maintenance_hours_month is not_measured, so no labor added (would be hours × $50/hr assumption if measured).). JSON: https://bot2botlabs.com/api/receipts/r-0001.json. Page: https://bot2botlabs.com/receipts/r-0001
- r-0002: PhotoCraft, the AI-built Photoshop clone: what it really costs to run (partially_verified, price_date 2026-10-19). Human hours: Not disclosed. Est. build tokens Not disclosed. Headline: $0.00/mo run vs $22.99/seat/mo. Key facts: On price date 2026-10-19, estimated monthly run cost is $0.00 (measured (local run on our box) + cited list prices) versus $22.99/seat/mo for the subscription it replaces. (2026-10-19 https://www.adobe.com/products/photoshop.html) Cash break-even is immediate at 1 seat: $0 vs $22.99/mo ($275.88/yr). The catch is fit, not cost: the builder rates it ~45% ready for real work (alpha), and it has none of Photoshop's generative AI features. (2026-10-19) Setup from README (release v0.10.0, Linux x86_64): Pass: 83.0 MB download, checksum OK, unpacked (159 MB) and running in about 5 s; no missing libraries. Measured 2026-10-19.. (2026-10-19) Core edit test set (our 15 fixed steps): 15/15 passed on release v0.10.0 and on source commit 5c89f16b: PNG↔PSD lossless, adjustment layers, resize, crop, rotate, text layer, drop shadow, JPG/WebP/TIFF, batch of 10, CMYK, 16-bit, bad file rejected without a crash. Measured 2026-10-19.. (2026-10-19) Build cost on price date 2026-10-19: Not disclosed human hours, Not disclosed estimated token dollars, billed not disclosed. (2026-10-19 https://github.com/storytold/photocraft/tree/5c89f16b60fb70195e874ff6d0ef1d3edba56e8a) Operator Y1 @1 user: $0.00 (measured_by_bot2bot_labs). User clone: free_self_host; real product @1 seat: $22.99/mo (price_date 2026-10-19). Break-even: Immediate (1 seat) — $0/mo clone vs $22.99/mo Photoshop: cheaper from the first seat and the first month ($275.88/yr per seat). Fit is the real cost: ~45% ready for real work (builder-stated), no generative AI features (measured). Verdict good_for: Solo users and small teams doing layered photo edits, adjustments, selections, type and batch/CLI automation who want $0 and offline files, and can live with alpha software. | skip_if: You rely on Generative Fill or Neural Filters, trade PSDs with Photoshop clients who need an exact match, or need vendor support, SOC 2 or an SLA. | main_risk: Alpha (~45% ready by the builder's own rating) and changing ~80 commits a day; a single commercial company behind it; legal questions cited, not settled. Confidence: 50% measured, 30% builder-stated, 20% unverified (10/20 key fields). Maintenance pulse (2026-10-19): last commit 2026-10-19T17:08:50Z, issues 1691, stars 68334, license MIT OR Apache-2.0. Data: Your own disk. The app made no internet connections in our test; no account or cloud storage. Export: yes. Price history (4 rows, re-check every 90d): Adobe Photoshop (individual, annual billed monthly) $22.99 on 2026-10-19; Adobe Photoshop (individual, month-to-month) $34.49 on 2026-10-19; Adobe Photoshop for teams (annual billed monthly) $37.99 on 2026-10-19; PhotoCraft v0.10.0 $0 on 2026-10-19. Labor-adjusted operator TCO Y1 @1: $0 (Same as cash TCO $0: maintenance hours not measured, so no labor added (would be hours × $50/hr × 12).). JSON: https://bot2botlabs.com/api/receipts/r-0002.json. Page: https://bot2botlabs.com/receipts/r-0002
- r-0003: free-slack-clone, the AI-built Slack clone: what it really costs to run (partially_verified, price_date 2026-10-21). Human hours: Not disclosed. Est. build tokens Not disclosed. Headline: $24.00/mo run vs $87.50/mo for 10 Slack seats. Key facts: On price date 2026-10-21, estimated monthly run cost is $24.00 (measured (local run on our box) + estimated hosting from cited list prices) versus $87.50/mo for 10 Slack seats for the subscription it replaces. (2026-10-21 https://slack.com/pricing) Cheaper than Slack Pro from 3 users (3 × $8.75 = $26.25 > $24). The catch: at commit fe9e13b the production build fails, so today it only runs in dev mode, and sign-in codes are not emailed. (2026-10-21) Setup from README (Node.js 22, Linux): Pass: git clone + npm install took about 7 s; `npm run dev` started embedded Postgres, ran 9 migrations, seeded a workspace and served the app in about 2 s. No Docker, no .env, no accounts. Measured 2026-10-21.. (2026-10-21) Builder's own smoke test (npm run smoke): 49/49 checks passed (README says "40+ assertions"). Measured 2026-10-21.. (2026-10-21) Build cost on price date 2026-10-21: Not disclosed human hours, Not disclosed estimated token dollars, billed not disclosed. (2026-10-21 https://github.com/guillaumeduhan/free-slack-clone/tree/fe9e13bb5e0e8a17f0305035e8562afa2dbcb885) Operator Y1 @1 user: $288.00 (estimate). User clone: free_self_host; real product @1 seat: $8.75/mo (price_date 2026-10-21). Break-even: From 3 users vs Slack Pro — $24/mo server (estimate) vs Slack Pro $8.75/user/mo: 2 users $17.50 < $24, 3 users $26.25 > $24. Paying Slack annually ($7.25) moves it to 4 users. Against Slack Free ($0) it never breaks even on cash; you would be paying for unlimited history and ownership. Verdict good_for: A small team that wants self-hosted chat (channels, threads, DMs, files, search, bots and webhooks) on one cheap server, and has someone who can fix a Next.js build and wire up email. | skip_if: You need huddles or video, Slack Connect, mobile push, SSO, Workflow Builder or AI, or nobody on the team can patch a Node app. Also skip if Slack Free's 90-day history is enough: that is $0. | main_risk: No working production build at this commit, so it only runs in dev mode today; sign-in codes are not emailed; one builder and no commits since 2026-10-12. Confidence: 68.4% measured, 5.3% builder-stated, 26.3% unverified (13/19 key fields). Maintenance pulse (2026-10-21): last commit 2026-10-12T18:53:53Z, issues 0, stars 0, license MIT. Data: Your own PostgreSQL database (schema "chat"), including uploaded files (stored in the database, 50 MB max each). Export: partial. Price history (5 rows, re-check every 90d): Slack Pro (paying monthly) $8.75 on 2026-10-21; Slack Pro (paying annually) $7.25 on 2026-10-21; Slack Business+ (paying monthly) $18 on 2026-10-21; Slack Pro (paying monthly) $8.75 on 2026-10-18; DigitalOcean Basic droplet 4 GiB / 2 vCPU $24 on 2026-10-21. Labor-adjusted operator TCO Y1 @1: $288 (Same as cash TCO $288: maintenance hours not measured, so no labor added (would be hours × $50/hr × 12).). JSON: https://bot2botlabs.com/api/receipts/r-0003.json. Page: https://bot2botlabs.com/receipts/r-0003
- r-0004: MailCheap, the AI-built beehiiv alternative: what it costs to self-host (partially_verified, price_date 2026-10-20). Human hours: Not disclosed. Est. build tokens Not disclosed. Headline: $0.64/mo run vs $49.00/mo (beehiiv Lite, up to 1,000 subscribers). Key facts: On price date 2026-10-20, estimated monthly run cost is $0.64 (estimate (cited list prices; live SES campaign bill and Atlas/Upstash usage not observed — see setup.blocked)) versus $49.00/mo (beehiiv Lite, up to 1,000 subscribers) for the subscription it replaces. (2026-10-20 https://www.beehiiv.com/pricing) At 1,000 subscribers and 4 sends/mo, self-host ~$0.64/mo vs beehiiv Lite $49/mo yearly (~$48/mo less before your time). Break-even on ops time depends on your hours (not measured); on infra alone it is cheaper from the first month. (2026-10-20) Local clone install (npm ci): 8.11 s wall-clock; 846 packages; Node 22.20.0. (2026-10-20 https://github.com/marclou/mailcheap/tree/bb82ec64f9f2157d73a2a2facb706b86013f25be) Unit + integration tests: 1032/1032 pass (Node 22.20.0, throwaway local mongod per tests). (2026-10-20 https://github.com/marclou/mailcheap/tree/bb82ec64f9f2157d73a2a2facb706b86013f25be) Build cost on price date 2026-10-20: Not disclosed human hours, Not disclosed estimated token dollars, billed not disclosed. (2026-10-20 https://github.com/marclou/mailcheap) Operator Y1 @1 user: $7.68 (estimate). User clone: MailCheap Cloud Starter (hosted) or self-host $0 + SES; real product (beehiiv Lite, up to 1,000 subscribers): $49.00/mo (price_date 2026-10-20). Break-even: Cheaper on infra from day one at 1,000 subscribers — ~$0.64/mo self-host (SES Essentials @ 4k emails) vs $49/mo beehiiv Lite yearly. Excludes your setup and maintenance hours. Verdict good_for: Operators who want newsletter send + archive + API/MCP and will run SES + Mongo themselves, especially around the 2026-11-01 beehiiv renewal change. | skip_if: You need beehiiv's ad network, podcasts, community, paid-subscription monetization, SOC 2 on the app vendor, or zero ops time. | main_risk: You own deliverability and AGPL obligations; campaign/QStash path not verified in this quick run and our SES test email landed in Spam without domain DKIM; SES Essentials is $0.16/1k for many new accounts vs builder's $0.10 claim. Confidence: 52.6% measured, 31.6% builder-stated, 15.8% unverified (10/19 key fields). Maintenance pulse (2026-10-20): last commit 2026-10-20T14:23:35Z, issues 13, stars 1265, license AGPL-3.0. Data: Self-host: your MongoDB + SES. Cloud: MailCheap's infra (not inspected). Export: yes. Price history (6 rows, re-check every 90d): beehiiv Lite, up to 1,000 subscribers (yearly billing) $49 on 2026-10-20; beehiiv Lite, up to 1,000 subscribers (monthly billing) $59 on 2026-10-20; beehiiv Lite, up to 10,000 subscribers (yearly billing) $119 on 2026-10-20; Amazon SES Essentials outbound (0–10M emails/mo) $0.16 on 2026-10-20; Amazon SES à la carte outbound $0.1 on 2026-10-20; MailCheap Cloud Starter (20,000 emails/mo) $19 on 2026-10-20. Labor-adjusted operator TCO Y1 @1: $7.68 (Same as tco_year_1 until hours measured.). JSON: https://bot2botlabs.com/api/receipts/r-0004.json. Page: https://bot2botlabs.com/receipts/r-0004
- r-0005: Minilytics, the AI-built DataFast alternative: what it costs to self-host (partially_verified, price_date 2026-10-26). Human hours: Not disclosed. Est. build tokens Not disclosed. Headline: $0.00/mo run vs $9.00/mo (DataFast Starter, 10k events/mo, 1 website). Key facts: On price date 2026-10-26, estimated monthly run cost is $0.00 (estimate (cited Cloudflare list prices; app, tests, burst and security measured locally; no Cloudflare deploy)) versus $9.00/mo (DataFast Starter, 10k events/mo, 1 website) for the subscription it replaces. (2026-10-26 https://datafa.st/) Free from day one at 1 site and 10k events/mo: $0/mo self-host vs $9/mo DataFast Starter (monthly), before your setup and upkeep time. At 100 sites the clone needs Workers Paid ($5/mo). (2026-10-26) Git history (build timeline, not hours): 16 commits, all authored "Claude" with an Opus 5.5 co-author trailer, from 2026-10-23 1:05 PM PT to 2026-10-26 5:07 AM PT. The first commit alone adds 10,212 lines across 49 files, so earlier work is not visible.. (2026-10-26 https://github.com/surendrapandar/minilytics/commits/main) Setup from README: Pass on Node 22.20.0 (install 9.5 s, migrate 3.4 s, build 6.4 s, local worker ready ~3 s). README says Node 20+, but Wrangler 4.152 refused Node 20.19.2.. (2026-10-26 https://github.com/surendrapandar/minilytics/blob/6f71720fa8ca996382d9d6e279161101e0f68ba1/README.md) Build cost on price date 2026-10-26: Not disclosed human hours, Not disclosed estimated token dollars, billed not disclosed. (2026-10-26 https://github.com/surendrapandar/minilytics) Operator Y1 @1 user: $0.00 (estimate). User clone: free_self_host (MIT) or hosted lifetime deal; real product (DataFast Starter, 10k events/mo, 1 website): $9.00/mo (price_date 2026-10-26). Break-even: Free from day one for 1 website at 10k events/mo — $0/mo self-host (Cloudflare free plan) vs $9/mo DataFast Starter monthly. 10 sites: $0 vs $39 (Growth 100k). 100 sites: $5 vs ~$295. Excludes your setup and upkeep hours. The hosted Minilytics lifetime deal ($29 once) equals ~3.2 months of DataFast Starter. Verdict good_for: Indie founders with a few sites who want DataFast-style revenue attribution (referrer → payment) on their own Cloudflare account for $0 and can wire payments through the API or Dodo. | skip_if: You need native Stripe/Shopify, team members, an analytics API or MCP, mobile apps, cookieless mode, imports, or a vendor that handles security for you. | main_risk: Three-day-old, single-maintainer project with one production critical (path looks unused), no security headers and non-revocable 30-day sessions; real Cloudflare behavior not tested by us. Confidence: 50% measured, 30% builder-stated, 20% unverified (10/20 key fields). Maintenance pulse (2026-10-26): last commit 2026-10-26T12:07:35Z, issues 5, stars 585, license MIT. Data: Self-host: your own Cloudflare D1 database. DataFast: its infrastructure (AWS, Vercel, MongoDB, Tinybird, ClickHouse and others per its DPA). Export: yes. Price history (8 rows, re-check every 90d): DataFast Starter, 10k events/mo, 1 website (monthly billing) $9 on 2026-10-26; DataFast Starter, 10k events/mo (yearly billing) $90 on 2026-10-26; DataFast Growth, 10k events/mo, 30 websites (monthly billing) $19 on 2026-10-26; DataFast Growth, 100k events/mo, 30 websites (monthly billing) $39 on 2026-10-26; DataFast Growth, 200k events/mo, 30 websites (monthly billing) $59 on 2026-10-26; Cloudflare Workers Paid (includes 10M requests/mo; D1 50M rows written/mo) $5 on 2026-10-26; Minilytics hosted lifetime deal (5 spots) $29 on 2026-10-26; Minilytics README sponsor slot (early price) $19 on 2026-10-26. Labor-adjusted operator TCO Y1 @1: $0 (Same as tco_year_1 until hours measured.). JSON: https://bot2botlabs.com/api/receipts/r-0005.json. Page: https://bot2botlabs.com/receipts/r-0005
- r-0006: GridCraft, the AI-built Excel clone: what it costs to run (partially_verified, price_date 2026-10-28). Human hours: Not disclosed. Est. build tokens Not disclosed. Headline: $0.00/mo run vs $9.99/mo (Microsoft 365 Personal, 1 person). Key facts: On price date 2026-10-28, estimated monthly run cost is $0.00 (measured (local run on our box) + cited list prices) versus $9.99/mo (Microsoft 365 Personal, 1 person) for the subscription it replaces. (2026-10-28 https://www.microsoft.com/en-us/microsoft-365/buy/compare-all-microsoft-365-products) Cash break-even is immediate: $0 vs $9.99/mo ($119.88/yr monthly, $99.99/yr yearly). The catch is fit, not cost: the builder rates it ~50% ready for real work (alpha), with no Copilot, co-authoring or mobile apps, and big sheets reopen cut short. (2026-10-28) Git history: Repo created 2026-10-07; 2,058 commits, 126 contributors, 12 releases in 21 days; 455 commits carry a Claude co-author line (GitHub search). v0.13.0 released 2026-10-28 9:10 AM PT.. (2026-10-28 https://github.com/storytold/gridcraft) Setup: release download: v0.13.0 Linux tarball 47,964,351 bytes, SHA-256 matches the release's SHA256SUMS.txt; download 2.6 s, extract and run; gridcraft-cli 0.13.0 lists 514 functions and 355 commands.. (2026-10-28 https://github.com/storytold/gridcraft/releases/tag/v0.13.0) Build cost on price date 2026-10-28: Not disclosed human hours, Not disclosed estimated token dollars, billed not disclosed. (2026-10-28 https://github.com/storytold/gridcraft) Operator Y1 @1 user: $0.00 (measured_by_bot2bot_labs). User clone: free_self_host; real product (Microsoft 365 Personal, 1 person): $9.99/mo (price_date 2026-10-28). Break-even: Free from day one vs Microsoft 365 Personal — $0/mo vs $9.99/mo = $119.88/yr per person paying monthly ($99.99/yr paying yearly). 10 people: $0 vs ~$235/mo on Business Standard with Copilot. Excludes your learning and support time, and the Word/Outlook/OneDrive/Copilot that Microsoft 365 also includes. Verdict good_for: Solo users and small teams who need everyday spreadsheets (formulas, lookups, PivotTables, charts) offline for $0, or who want agents to drive a spreadsheet through its CLI or MCP server. | skip_if: You need Copilot, co-authoring, mobile apps, VBA or Power Query, very large workbooks, or files that must round-trip perfectly with colleagues on Excel. | main_risk: Alpha software three weeks old: sheets over ~500k cells reopen cut short (open issue #2526), and fit with real Excel files is unproven outside our 350-file test. Confidence: 69.6% measured, 13% builder-stated, 17.4% unverified (16/23 key fields). Maintenance pulse (2026-10-28): last commit 2026-10-28T16:10:32Z, issues 733, stars 6717, license MIT OR Apache-2.0. Data: Your own disk. No account or cloud storage; the desktop app only contacts GitHub for update checks. Export: yes. Price history (7 rows, re-check every 90d): Microsoft 365 Personal (monthly) $9.99 on 2026-10-19; Microsoft 365 Personal (monthly) $9.99 on 2026-10-28; Microsoft 365 Personal (yearly) $99.99 on 2026-10-28; Microsoft 365 Family, up to 6 people (monthly) $12.99 on 2026-10-28; Microsoft 365 Family, up to 6 people (yearly) $129.99 on 2026-10-28; Microsoft 365 Business Standard with Copilot (paid yearly) $23.5 on 2026-10-28; GridCraft v0.13.0 $0 on 2026-10-28. Labor-adjusted operator TCO Y1 @1: $0 (Same as cash TCO $0: maintenance hours not measured, so no labor added (would be hours × $50/hr × 12).). JSON: https://bot2botlabs.com/api/receipts/r-0006.json. Page: https://bot2botlabs.com/receipts/r-0006
- r-0007: Everything Accounting, the open-source QuickBooks clone: what it costs to run (partially_verified, price_date 2026-10-30). Human hours: Not disclosed. Est. build tokens Not disclosed. Headline: $0.00/mo run vs $38.00/mo (QuickBooks Online Simple Start, 1 user). Key facts: On price date 2026-10-30, estimated monthly run cost is $0.00 (measured (local run on our box) + cited list prices) versus $38.00/mo (QuickBooks Online Simple Start, 1 user) for the subscription it replaces. (2026-10-30 https://quickbooks.intuit.com/pricing/) Cash break-even is immediate: $0 (or $10/mo on Puter Basic) vs $38/mo = $336–$456/yr per business at list price. The catch is fit: no bank feeds, online payments, multiple users or accountant access, and no audits behind it. (2026-10-30) How it was built (git history): The Everything repo has 1,057 'Co-Authored-By: Claude Opus 5.5' lines across its 255 commits. The Accounting app itself arrived in one commit (5f183ee, 2026-10-13, ~21,000 lines) with no co-author line, so AI authorship of this app is not shown in its own history.. (2026-10-30 https://github.com/HeyPuter/everything/commit/5f183ee1) Project activity: Repo created 2026-10-13; 5,900 stars, 641 forks, 58 open issues, 3 contributors (jelveh 248 commits). src/accounting unchanged since 2026-10-13.. (2026-10-30 https://github.com/HeyPuter/everything) Build cost on price date 2026-10-30: Not disclosed human hours, Not disclosed estimated token dollars, billed not disclosed. (2026-10-30 https://github.com/HeyPuter/everything) Operator Y1 @1 user: $0.00 (measured_by_bot2bot_labs). User clone: free_self_host; real product (QuickBooks Online Simple Start, 1 user): $38.00/mo (price_date 2026-10-30). Break-even: Free from day one vs QuickBooks Simple Start — $0/mo on Puter Free vs $38/mo list = $456/yr per business ($399 in year 1 with the 3-month promo). On Puter Basic ($10/mo) the saving is $28/mo = $336/yr. Versus QuickBooks Free or Wave Starter the cash saving is $0. Excludes your time and the bank feeds, payments and multi-user access you give up. Verdict good_for: Owners and bookkeepers who want real double-entry books (invoices, bills, reconcile, P&L and balance sheet that tie out) for $0, and are fine entering bank lines from statement files. | skip_if: You need bank feeds, online invoice payments, payroll, inventory, more than one login per company, or an accountant who works in QuickBooks. | main_risk: Your books depend on Puter: sign-in, storage and its unpublished free allowance. The app has no audits, no security headers in its hosting config, and no changes since 2026-10-13. Confidence: 73.9% measured, 8.7% builder-stated, 17.4% unverified (17/23 key fields). Maintenance pulse (2026-10-30): last commit 2026-10-13T16:38:36Z, issues 58, stars 5900, license MIT. Data: Your own Puter account (key-value store and drive) plus a cache in your browser. No server of the builder's. Export: yes. Price history (8 rows, re-check every 90d): QuickBooks Online Simple Start (list) $38 on 2026-10-30; QuickBooks Online Essentials (list, 3 users) $85 on 2026-10-30; QuickBooks Online Plus (list, 5 users) $140 on 2026-10-30; QuickBooks Online Advanced (list, 25 users) $340 on 2026-10-30; Xero Early (US) $27 on 2026-10-30; Wave Pro $19 on 2026-10-30; Puter Basic $10 on 2026-10-30; Everything Accounting @ b5d2584 $0 on 2026-10-30. Labor-adjusted operator TCO Y1 @1: $0 (Same as cash TCO $0: maintenance hours not measured, so no labor added (would be hours × $50/hr × 12).). JSON: https://bot2botlabs.com/api/receipts/r-0007.json. Page: https://bot2botlabs.com/receipts/r-0007

## Price snapshots (2026-10-16)
- Anthropic Claude Sonnet 4 — input tokens: 3 (USD per 1M input tokens) checked 2026-10-16 https://www.anthropic.com/pricing#api
- Anthropic Claude Sonnet 4 — output tokens: 15 (USD per 1M output tokens) checked 2026-10-16 https://www.anthropic.com/pricing#api
- OpenAI GPT-4o mini — input tokens: 0.15 (USD per 1M input tokens) checked 2026-10-16 https://openai.com/api/pricing/
- OpenAI GPT-4o mini — output tokens: 0.6 (USD per 1M output tokens) checked 2026-10-16 https://openai.com/api/pricing/
- Vercel Hobby: 0 (USD per month (list)) checked 2026-10-16 https://vercel.com/pricing
- Vercel Pro: 20 (USD per month) checked 2026-10-16 https://vercel.com/pricing
- Resend Free tier: 0 (USD per month) checked 2026-10-16 https://resend.com/pricing
- Resend Pro: 20 (USD per month) checked 2026-10-16 https://resend.com/pricing

See https://bot2botlabs.com/llms.txt for canonical short summary.