AI Summary
About
Helicone is an open-source LLM observability platform and AI gateway. With a one-line proxy integration (edge-deployed on Cloudflare Workers) or async logging, it captures every LLM request to give developers cost, latency, and quality visibility across providers like OpenAI, Anthropic, Azure, OpenRouter, Together AI, and LiteLLM. A YC W23 company, Helicone became one of the most popular open-source projects in the LLM tooling space (Apache 2.0), hit $1M ARR in June 2024 with roughly a five-person team, and raised a $5M seed at a ~$25M valuation in September 2024 (led by Y Combinator and Village Global, with FundersClub).
In March 2026, Mintlify acquired Helicone and moved the hosted product into maintenance mode — security patches, bug fixes, and new-model support continue, but active feature development has stopped. Founders Justin Torre and Cole Gottdank joined Mintlify. The open-source repository remains available.
For the most current information, visit Helicone.
Pricing summary : How Helicone’s pricing model works
Helicone bills per organization across three self-serve tiers plus a quoted Enterprise plan. Hobby is free (10,000 requests/month, 1 GB storage, 1 seat, 7-day retention). Pro is $79/month and unlocks unlimited seats, alerts, reports, and HQL (the query language), with 1-month retention. Team is $799/month, adds a second-through-fifth organization, SOC-2 & HIPAA compliance, 3-month retention, and a dedicated Slack channel. Enterprise is custom-quoted and layers on SAML SSO, on-prem deployment, a custom MSA, and bulk discounts.
The twist: Pro and Team are not all-you-can-eat. Each includes the same 10,000 free requests and 1 GB of storage as Hobby, after which usage-based overage kicks in on logs (requests) and storage. Helicone’s on-page calculator estimates roughly $0.97/month for 10,000 requests with light storage — so a small production workload pays the flat $79 plus a few dollars of usage.
What makes this different: Helicone is genuinely open-source (Apache 2.0) and fully self-hostable for free with no request cap — the paid tiers buy hosted convenience, compliance, retention, and team features rather than the core capability. The free-tier request allowance has swung dramatically over time (100K → 50K → 100K → 10K), and the headline Pro plan has cycled through pure usage, a flat monthly fee, per-seat, and back to the current flat per-org $79 (see Pricing evolution).
Pricing by product
| Tier | Price | Included | Key mechanics |
|---|---|---|---|
| Hobby | Free | 10K requests/mo, 1 GB storage, 1 seat, 7-day retention | Free forever, no card; community support |
| Pro | $79/mo per org | 10K req + 1 GB free, then usage-based; unlimited seats, 1-month retention | Alerts, reports, HQL; 7-day free trial |
| Team | $799/mo per org | 10K req + 1 GB free, then usage-based; 5 orgs, 3-month retention | SOC-2 & HIPAA, dedicated Slack; 7-day free trial |
| Enterprise | Custom | Unlimited, forever retention | SAML SSO, on-prem, custom MSA, bulk discounts |
| Self-host (OSS) | Free | No request cap | Apache 2.0; you run the infrastructure |
Sales motions across products: self-serve PLG for Hobby/Pro/Team (instant signup, 7-day trial on paid), open-source self-host, and sales-led for Enterprise. Targeted discounts: 50% off year one for startups (under 2 yrs, under $5M raised), free for students/educators, $100 first-year credit for open-source companies, and discounts for non-profits.
Hidden costs : What Helicone users actually pay
The flat monthly price is only the floor. Real Helicone bills are shaped by log (request) volume and storage above the 10K/1 GB included allotment, both metered as usage-based overage. Helicone does not publish per-unit overage rates on the pricing page — only an interactive calculator — so the figures below are illustrative estimates derived from that calculator; treat exact overage rates as quote-dependent.
| Line item | Monthly cost (illustrative) |
|---|---|
| Pro base plan | $79 |
| Usage overage — light workload (calculator shows ~$0.97 for 10K req + light storage) | ~$1–$10 |
| Extended data retention / large storage | usage-based (varies) |
| Estimated total (small production app) | ~$80–$90 |
Other things to budget for: retention is short on lower tiers (7 days on Hobby, 1 month on Pro, 3 months on Team) so long-horizon analytics effectively require Team or Enterprise; SOC-2/HIPAA compliance is gated to Team ($799) and above; and SAML SSO / on-prem are Enterprise-only.
Want to estimate your own Helicone bill? Use the Helicone pricing calculator to model your costs based on usage patterns.
Pricing evolution : Helicone pricing history and changes
Cadence
| Period | Price changes | Product / SKU additions | Notes |
|---|---|---|---|
| 2023 | Launch | Basic Flex (pure usage) | 100K free req, then $1/10K req |
| 2024 H1 | Pro $80 | Free / Pro / Custom tiers | Free cut to 50K req logs |
| 2024 H2 | — | Growth (usage slider) | Free raised back to 100K |
| 2025 H1 | Pro $20/seat | Per-seat model | Free cut to 10K req |
| 2026 | Pro $20→$79, Team added | Hobby / Pro / Team / Enterprise | Flat per-org; +Team $799 |
Tracked range: 2023–present, via Wayback Machine snapshots (2023-06, 2024-03, 2024-08, 2025-01) and a live 2026-06-09 capture.
Notable changes
- 2023-06 — Launches with Basic Flex: 100K free requests/month, then $1.00 per 10K requests. Pure usage, no seats, no subscription. Enterprise via contact-sales.
- 2024-03 — Restructures to Free ($0, 50K logs) / Pro ($80/mo, 500K logs) / Custom. First move toward included-allotment tiers.
- 2024-08 — Free tier raised to 100K requests; introduces a Growth plan with an interactive usage slider (100K–10M requests).
- 2025-01 — Free tier cut to 10K requests; Pro becomes $20/user/month plus usage-based overage.
- 2026 — Pro moves off per-seat to flat $79/month per org; new Team tier at $799/month (5 orgs, SOC-2/HIPAA). Hobby stays free at 10K requests + 1 GB.
- 2026-03 — Mintlify acquires Helicone; hosted product enters maintenance mode (bug fixes, security, new models only).
What’s unique : Helicone’s distinctive pricing mechanics
1. Truly free via open source. Unlike most “open-core” observability tools, Helicone’s full platform is Apache 2.0 and self-hostable with no request cap. The paid SaaS sells hosting, compliance, retention, and team management — not the core capability. This caps how much pricing power Helicone can exert.
2. Flat subscription floor + usage overage on logs/storage. Pro ($79) and Team ($799) aren’t all-you-can-eat: each includes only 10K requests + 1 GB, then meters overage. This hybrid keeps a predictable monthly anchor while still scaling with heavy log volume — and avoids the per-seat penalty that punishes large teams.
3. Per-org, not per-seat (again). Helicone tried per-seat ($20/user) in 2025 and reverted to flat per-org pricing in 2026 with unlimited seats on Pro. Removing the seat tax is unusual in observability and favors collaborative teams.
Strengths & weaknesses
| Strengths | Weaknesses |
|---|---|
| Generous, genuinely-open-source free path (self-host, no cap) | Hosted product in maintenance mode post-acquisition — feature risk |
| Predictable flat floor ($79/$799) with usage only on overage | Per-unit overage rates not published — only a calculator |
| Unlimited seats on Pro (no per-seat tax) | Short retention on lower tiers (7d/1mo) forces upgrades |
| One-line proxy integration across 100+ models/providers | Proxy is a single point of failure for live LLM calls |
| Compliance (SOC-2/HIPAA) available without full Enterprise | Cost metadata for brand-new models can lag provider releases |
Billing UX : Helicone billing controls and transparency
- Billing controls — Self-serve upgrade/downgrade between Hobby, Pro, and Team; a 7-day free trial on paid plans; no credit card needed for Hobby. Plan switching and cancellation are self-service.
- Usage visibility — The pricing page ships an interactive cost calculator (requests × tokens/request + storage) showing a live estimate, plus in-product cost dashboards, alerts, and reports (alerts/reports gated to Pro+). This is strong forward visibility, partly offset by the lack of published per-unit overage rates.
- Payment options — Standard card-based self-serve checkout for Hobby/Pro/Team; Enterprise is invoiced under a custom MSA with bulk cloud discounts. Targeted discount programs (startups, students, open-source, non-profits) require an application.
Strategic wins : Why Helicone’s pricing decisions worked
1. Open-source as the top-of-funnel
Leading with an Apache-2.0, one-line-integration product turned GitHub stars and HN launches into a self-serve funnel — Helicone reached $1M ARR with ~5 people. The free self-host path removes adoption friction; the hosted tiers monetize the teams that don’t want to run infrastructure. See how AI companies structure pricing.
2. Reverting from per-seat to per-org
The 2025 experiment with $20/seat created friction for collaborative teams. Going back to a flat $79/org with unlimited seats in 2026 removed the seat tax and made expansion painless — a deliberate choice to grow accounts by usage, not headcount. Related: outcome-based pricing trends.
3. A flat floor with usage overage
Anchoring on $79/$799 with only overage on logs/storage gave buyers a predictable line item while preserving upside on heavy workloads — a cleaner mental model than the pure $1/10K usage of 2023. See choosing the right usage metric.
Areas to improve : Gaps in Helicone’s pricing approach
1. Publish the overage rates
The pricing page shows a calculator but never states the per-log or per-GB price, so buyers can’t reason about cost at scale without modeling each scenario. Publishing the unit rates would close a real transparency gap. See bill shock and cost unpredictability.
2. Whiplash on the free tier
The free request allowance swung 100K → 50K → 100K → 10K in under two years. Frequent, large changes to the free anchor erode trust and make capacity planning hard for the developers who form the funnel.
3. Post-acquisition roadmap uncertainty
Maintenance mode under Mintlify means new models and ecosystem changes may not get fast fixes. For a fast-moving category, that’s a pricing-relevant risk: buyers are paying a subscription for a product whose feature velocity has stopped.
Key takeaways
- Open source can be the funnel, not the giveaway. Helicone monetized hosting, compliance, and retention on top of a free Apache-2.0 core — and still hit $1M ARR with five people.
- Per-seat pricing isn’t always right for collaborative tools. Helicone tried $20/seat and reverted to flat per-org with unlimited seats.
- A flat floor + usage overage beats pure usage for predictability. The 2023 “$1 per 10K requests” model became a flat $79 with overage as buyers matured.
- Free-tier stability matters. Repeated swings in the free request cap (100K↔10K) undercut trust with the very developers driving adoption.
- Consolidation is reshaping LLM observability. Mintlify’s acquisition of Helicone signals that standalone observability may struggle to stay defensible — a pricing-power lesson for the whole category.
UBP implications
- Hybrid (flat + overage) is the maturing default for observability. A predictable subscription anchor with usage only on the metered dimensions (logs, storage) balances buyer predictability against vendor upside.
- Pick a value metric your buyer already counts. Requests/logs map directly to what developers instrument, making the meter intuitive — but the unit rate must be published to be trusted.
- Open source changes pricing power. When the core is free to self-host, paid tiers must sell convenience, compliance, and retention — not the capability itself. See usage-based pricing strategy.
Sources
- Helicone pricing page (live capture, accessed 2026-06-09)
- Helicone GitHub (Apache 2.0, YC W23) (accessed 2026-06-09)
- Helicone: joining Mintlify (accessed 2026-06-09)
- Mintlify acquires Helicone (accessed 2026-06-09)
- Y Combinator — Helicone (accessed 2026-06-09)
- Helicone cost tracking docs (accessed 2026-06-09)
- Wayback Machine snapshots: helicone.ai/pricing — 2023-06, 2024-03, 2024-08, 2025-01 (accessed 2026-06-09)
Bottom line
Helicone is an open-source (Apache 2.0), YC W23 LLM observability platform and AI gateway that turned a one-line proxy and a free self-host path into a self-serve funnel reaching $1M ARR with five people. Hosted pricing is per-organization — free Hobby (10K requests), $79 Pro and $799 Team (each 10K requests + 1 GB included, then usage-based overage on logs/storage), and quoted Enterprise. Its pricing has cycled through pure usage, per-seat, and flat per-org models; after Mintlify’s March 2026 acquisition the hosted product is in maintenance mode. Browse the pricing blueprint for more fully-researched company profiles.
Want to compare Helicone against other LLM observability companies like LangSmith, Langfuse, or Braintrust? Browse the pricing blueprint.
Pricing timeline : Major events on a vertical axis
Each milestone below corresponds to a public pricing change, product launch, or material adjustment. Major events use a filled marker; minor adjustments use a faded one.
Flat per-org Pro $79 / Team $799 + usage
Current structure: Hobby (free, 10K req, 1 GB), Pro $79/mo and Team $799/mo (each 10K req + 1 GB free, then usage-based overage on logs/storage), and Enterprise. Pro moved off per-seat to flat per-org and rose to $79; a $799 Team tier (5 orgs, SOC-2/HIPAA) was added.
Per-seat Pro at $20/user + free tier cut to 10K
Free monthly requests cut from 100K to 10K. Pro became $20 per user/month plus usage-based overage. Three-tier Free / Pro / Enterprise structure.
Free 100K + Growth usage-slider
Free tier raised to 100K requests/month; introduced a 'Growth' plan with an interactive usage slider (100k–10m requests) sitting between Free and Enterprise.
Tiered Free / Pro / Custom restructure
Moved to a 3-tier comparison: Free ($0, 50K request logs/mo), Pro ($80/mo, 500K logs/mo) and Custom/Enterprise. Shift from pure usage toward included-allotment tiers.
Launch pricing — Basic Flex pure usage
Helicone (YC W23) launches with a single self-serve 'Basic Flex' plan: 100,000 free requests/month, then $1.00 per additional 10,000 requests. Plus a contact-sales Enterprise tier.
- · Helicone is a Y Combinator W23 company and one of the most-starred open-source LLM observability projects on GitHub (Apache 2.0).
- · It hit $1M ARR in June 2024 with a team of about five people, then raised a $5M seed at a ~$25M valuation in September 2024 (led by YC and Village Global).
- · Helicone's launch plan in 2023 was pure usage: 100K free requests/month, then exactly $1.00 per 10K requests — no seats, no subscription.
Questions & answers
- What is Helicone's pricing model?
- Helicone uses per-organization subscriptions — Hobby (free), Pro ($79/mo), and Team ($799/mo) — each with usage-based overages on logs and storage above the included 10,000 requests and 1 GB. Enterprise is custom-quoted, and the open-source version is free to self-host under Apache 2.0.
- Does Helicone offer a free tier?
- Yes. The Hobby plan is free forever with 10,000 requests/month, 1 GB storage, 1 seat and 7-day data retention — no credit card required. You can also self-host the open-source platform for free with no request cap.
- How much does Helicone cost per month?
- Pro is $79/month and Team is $799/month, both billed per organization. Each includes 10K requests and 1 GB storage, after which usage-based overage applies. Helicone's calculator shows roughly $0.97/month for 10K requests with light storage; Enterprise pricing is custom.
- Is Helicone pricing usage-based or subscription?
- It is a hybrid. You pay a flat per-organization subscription (Pro $79, Team $799) plus usage-based overage for logs and storage consumed beyond the included allotment. The earliest 2023 plan was pure usage ($1 per 10K requests after 100K free).
- Is Helicone still being developed after the Mintlify acquisition?
- Mintlify acquired Helicone in March 2026 and moved the hosted product into maintenance mode — security patches, bug fixes and new-model support continue, but no new feature development. The open-source repo remains under Apache 2.0.