Skip to main content
Beton
posthog_credits_token posthog_credits_token_mvp_v1

Credits / token-based

Prepaid credit or token systems where burn and top-up patterns matter more than traditional funnels. Low-balance warnings into purchases are the key positive signal.

Value metric

Credits consumed, tokens used, compute units

Success event

credits_purchased

entity_type: account

Scale

  • 3,000 accounts
  • 8 users/account (mean)
  • 16 sessions/user (mean)
  • 300 days of history

Research metrics proxied

  • — Burn rate
  • — Days-to-zero
  • — Top-up frequency
  • — Auto-refill adoption

Signal paths

Positive paths end in credits_purchased. Negative paths do not. Every generated event_id belonging to a path is recorded in ground_truth.json.

Positive signals (2)

warning_to_purchase ×850
low_balance_warning → credits_purchased
cohorts: high_intent, medium_intent, power_user
auto_refill_conversion ×450
low_balance_warning → auto_refill_triggered → credits_purchased
cohorts: high_intent, power_user

Negative signals (2)

burn_to_warning_only ×650
credits_used → low_balance_warning
cohorts: medium_intent, low_intent, lurker
repeated_warning_decline ×350
low_balance_warning → low_balance_warning
cohorts: low_intent, noisy_bot_like

Generate this dataset

Config file: configs/posthog_credits_token_mvp.yaml

Quickstart 
# Dockerized Postgres (recommended for inspection)
docker compose up -d

uv run dryfit \
  -c configs/posthog_credits_token_mvp.yaml \
  --dsn postgresql://dryfit_writer:dryfit_writer@127.0.0.1:54329/dryfit \
  --print-summary

# Or local Postgres
./scripts/generate-local -c configs/posthog_credits_token_mvp.yaml --print-summary

Full setup instructions are in the repo's README — including local Postgres, Grafana inspection, and dataset restore.

Noise parameters

DryFit injects realistic noise on top of the generated signal paths. These probabilities are per-event. Noise never touches rows referenced by ground_truth.json — your scoring logic can trust the truth file is exact.

missing event probability
5.0%
duplicate event probability
2.0%
out of order probability
3.0%
null property probability
3.0%
anonymous actor probability
1.0%
weird property probability
2.0%

Other scenarios

Combined coverage (all models)

→ upgrade_clicked

Union of event types across all business models

Contact / record-based SaaS

→ contact_created

Contacts, leads, subscribers, accounts managed

Event-volume SaaS

→ custom_event_tracked

Events tracked, data points ingested, log lines

Feature-gated (tiered) SaaS

→ upgrade_clicked

Plan tier / feature access level
Browse all scenarios →

Benchmark your detector against Credits / token-based

Clone the repo, run the config, check your agent's output against ground_truth.json.

View on GitHub