Tag
22 posts tagged.
Modal bills GPU compute by the second with scale-to-zero. We decode the 2026 per-GPU rate card, the keep-warm idle trap, three real 30-day bills, and the ~73% utilization crossover versus a rented pod.
Braintrust bills by evaluation scores (rows x scorers x runs), not seats. We decode the meter, find the ~199,000-score Pro crossover, work three 30-day bills, and show a one-lever fix that cut a bill 71%.
LaunchDarkly's 2026 pricing no longer charges per seat. The paid Foundation plan meters two lines: $10 per service connection per month plus $8.33 per 1,000 client-side monthly active users (MAU). The MAU meter bills your highest-volume context kind, so a 500,000-device app pays $4,165 a month on MAU alone even with far fewer logged-in users. A worked mid-size bill lands near $4,565 a month, about 91% driven by the MAU meter. Seats are free; the meter that hurts is client-side MAU.
Tinybird's 2026 pricing bills vCPU-seconds, not queries or storage. One always-on vCPU is $518.40/month. Here is the worked 30-day bill and the only lever that cuts it.
As of July 2026, Knock's notification API charges per message, where a message is one delivery to one user on one channel. The Starter plan is $250 per month for 50,000 messages, then $0.005 each. Because Knock meters per channel, a single notification sent to email, push, and in-app counts as three messages, so a 5,000-user app fanning three channels can run about $750 per month. The same alerts cost roughly $250 on Courier's per-notification meter and about $54 on Novu's per-run Pro plan. Match the meter to your channel fan-out before you pick a vendor.
As of July 2026, PostHog has no fixed subscription. It is free to start with no credit card, and each of its products bills separately once you pass a generous monthly free tier: 1M product-analytics events, 5K session replays, 1M feature-flag requests, 100K error-tracking exceptions, and more. Product analytics starts at $0.00005 per anonymous event and gets cheaper per event as volume rises. The catch nobody flags: identified events can cost up to 4x more. Most side projects pay $0; a growing SaaS usually lands between roughly $50 and a few hundred dollars a month; the five-figure horror stories come from high-volume, autocapture-heavy, all-identified setups with no billing limit set.
As of July 2026, Upstash Redis is free up to 256 MB and 500K commands per month; pay-as-you-go bills $0.20 per 100K commands plus $0.25/GB stored, and fixed plans run $10 to $1,500/month. Because Upstash charges per command, not per server, request volume sets your bill, not data size. Above ~5M commands/month a flat $10 fixed plan usually beats pay-as-you-go.
Mux bills video in three meters, not one plan price. Here is the real 2026 arithmetic: worked 30-day bills for three app sizes, the resolution multiplier that quietly quadruples costs, and four levers that lower the bill.
Pinecone runs on usage-based serverless pricing in 2026: a free Starter tier, a $20/month flat Builder plan, and a Standard plan with a $50/month minimum. You pay $16 per million read units, $4 per million write units, and $0.33 per GB of storage each month. The catch most teams miss: a query costs one read unit per GB of namespace size, not per query, so your bill is set by how you shape your namespaces, not how many searches you run. A hobby RAG app fits the free tier at $0; a 20 GB single-corpus production app runs about $110/month; a 50-million-vector multi-tenant app with per-tenant namespaces lands near $75.
Cloudflare Workers bills CPU time, not duration, which inverts everything you know from AWS Lambda. Here is the 2026 pricing, three worked monthly bills, and the 3-millisecond breakpoint that decides your cost.
As of July 2026, SendGrid no longer has a permanent free plan: the floor is a 60-day trial then Essentials at $19.95/mo (50,000 emails). Pro starts at $89.95/mo (100,000 emails plus one dedicated IP). Your real bill is the base price plus per-email overage ($0.0013 on the 50K plan, $0.0009 on 100K, $0.0011 on entry Pro). A 10K/mo sender pays a flat $19.95; 200K/mo on Pro runs about $199.95 if you eat overage.
Neon bills Postgres by the CU-hour with scale-to-zero. Here is the 2026 rate card, the always-on trap that 5x-es your bill, three worked 30-day bills, and the crossover versus a fixed instance.
