The two most common tools developers reach for when they need LLM cost control are Helicone and Portkey. Both are marketed as solutions to the same problem: you're spending too much on AI API calls and you need visibility and control.
The comparison gets interesting fast. Helicone was acquired by Mintlify on March 3, 2026, and is now in maintenance mode. Portkey announced a pending acquisition by Palo Alto Networks on April 30, 2026. Both products are in the middle of ownership transitions at the same time — and neither one actually enforces a hard budget cap at a self-serve price point.
This post covers what each tool does, what it doesn't do, and what that means if you need enforcement rather than observation. If you want a broader comparison across more tools, see our post on LiteLLM alternatives in 2026.
What Helicone Does
Helicone is an LLM observability platform. The core product is a proxy that sits between your application and the model provider — you change your baseURL to a Helicone endpoint, and Helicone starts logging every request and response. The setup is roughly 2 minutes.
What you get from that logging:
- Cost attribution by user, session, or custom property
- Latency and error rate tracking
- Request replay and debugging
- Rate limiting by request count (not by spend)
The pricing before the Mintlify acquisition:
| Tier | Price | Requests/mo | Retention |
|---|---|---|---|
| Hobby | Free | 10k | 7 days |
| Pro | $79/mo | 10k + overage | 1 month |
| Team | $799/mo | Unlimited | 3 months |
The $720/month jump from Pro to Team — with nothing in between — was a persistent complaint in the community. There is no $150/mo or $249/mo option that gets you meaningful retention or scale.
The Maintenance Mode Problem
Helicone's announcement of the Mintlify acquisition states: "Helicone's services will remain live for the foreseeable future in maintenance mode." Security updates, new model support, and bug fixes continue shipping — but active feature development has stopped.
This is a meaningful statement for production teams. Maintenance mode means:
- New feature requests go unanswered
- Roadmap items are stalled pending Mintlify's integration decisions
- Long-term pricing and SLA commitments are uncertain
For teams evaluating Helicone today, the acquisition timeline introduces risk that didn't exist three months ago. A tool in maintenance mode is appropriate for stable, low-change use cases. It's a harder sell for teams that need active development against their requirements.
What Helicone Doesn't Do
Helicone is explicitly an observability tool. It shows you what you spent. It does not stop you from spending it.
- No hard budget caps that block calls before they fire
- No per-user spending enforcement
- No passthrough billing to end-users
- Rate limits are by request count, not by dollar amount
If your user spends $200 when their limit should be $50, Helicone shows you that in the dashboard. The $200 is already gone.
What Portkey Does
Portkey is positioned as a full production stack: gateway, observability, guardrails, and governance in one platform. The feature set is substantially wider than Helicone's.
The current pricing:
| Tier | Price | Recorded logs/mo |
|---|---|---|
| Developer | Free | 10k |
| Production | $49/mo | 100k (+ $9/100k overage) |
| Enterprise | Custom | 10M+ |
Portkey includes routing features that Helicone doesn't: automatic fallbacks, load balancing across providers, semantic caching, and deterministic guardrails. For teams building with multiple model providers, this is a meaningful capability difference.
On April 30, 2026, Palo Alto Networks announced its intent to acquire Portkey. The announcement is confirmed on Palo Alto Networks' press page and visible in Portkey's own site header. The deal is pending close in Q4 FY2026 and positions Portkey's gateway and guardrails within Palo Alto's AI security stack.
The Enforcement Lock
The critical limitation for cost control use cases: budget enforcement is an Enterprise-only feature.
Portkey's pricing page lists "Budget & rate limits" under Enterprise. This is not a marketing footnote — it is the actual product gating. If you're on Developer (free) or Production ($49/mo), you have visibility into your spend. You do not have enforcement.
What this means in practice:
- A Production ($49/mo) user who wants to cap spending per customer cannot do it natively
- Enforcement requires Enterprise pricing — custom, sales-gated, likely $500+/mo
- There is a $450+ pricing cliff between "see your costs" and "enforce your costs"
For solo founders and small teams — exactly the people most exposed to runaway agent costs — Portkey offers observability without the enforcement they need.
Guardrails vs. Budget Enforcement
Portkey has a guardrails feature that is separate from budget enforcement. Guardrails validate prompt/response content — checking for PII, toxic content, format violations. This is content governance, not cost governance.
Budget enforcement is a distinct mechanism: it checks remaining spend against a threshold before firing the API call. A guardrail that blocks toxic output doesn't help when your agent hits an edge case and loops 800 times at 4 cents per call.
Helicone vs Portkey: Feature Comparison
| Capability | Helicone | Portkey |
|---|---|---|
| Cost visibility | ✓ All tiers | ✓ All tiers |
| Request logging | ✓ All tiers | ✓ All tiers |
| Per-user attribution | ✓ Pro+ ($79) | ✓ All tiers |
| Hard budget caps | ✗ Never | Enterprise only |
| Pre-flight enforcement | ✗ Never | Enterprise only |
| Per-user billing passthrough | ✗ | ✗ |
| Multi-provider routing | Limited | ✓ All tiers |
| Acquisition status | Mintlify (maintenance mode) | Palo Alto Networks (pending) |
Both tools are strong at observability. Neither provides self-serve enforcement.
Why Both Tools Are Observability First
The observability-first design is a deliberate choice, not an oversight. These tools were built to answer: "what did I spend?" The market consensus when both tools launched was that visibility was the missing piece — if you could see the costs, you could control them manually.
The problem that has emerged since: visibility doesn't stop the call. By the time Helicone or Portkey logs the token count, the request has already been sent, the provider has already processed it, and the cost is already on your invoice.
For deterministic workloads — a chatbot with predictable input length — this is workable. You look at the dashboard, notice costs are trending high, and increase the context window or tighten the prompt. You're always catching up to yesterday's numbers.
For agent workloads, the model breaks down. An agent runs autonomously. It can enter a retry loop at 3am. It can process a 200k-token document because the user uploaded the wrong file. It can call an expensive model 400 times in 8 minutes. None of these are visible until after the fact — and after the fact is too late.
What Pre-Flight Enforcement Looks Like
Pre-flight enforcement checks the estimated cost of each call before it fires. The flow:
- Your application calls the enforcement gateway with a model and input
- The gateway estimates the cost (input tokens × model rate)
- It checks that cost against the remaining budget for that user/project
- If remaining budget ≥ estimated cost → call proceeds
- If remaining budget < estimated cost → call is rejected with a structured error
The application receives the rejection before any tokens are consumed. The user hasn't spent anything. You can return a friendly error ("you've reached your monthly limit"), offer an upgrade path, or fail gracefully.
For a technical walkthrough of the three enforcement approaches and how to implement pre-flight checking yourself, see How to Set a Hard Budget Cap on LLM API Calls.
Portkey does this at Enterprise. Helicone does not do it at any tier.
The Acquisition Factor
Both tools are in ownership transitions. This is a real consideration for anyone building on top of them:
Helicone + Mintlify: Mintlify builds documentation tools — a complementary but different product category. The joint announcement explicitly points users toward migration alternatives and confirms that new feature development has stopped. The direction of any future Mintlify integration is unclear.
Portkey + Palo Alto Networks: Palo Alto is an enterprise security vendor with a $100B+ market cap. The acquisition announcement positions Portkey's technology within an enterprise security stack. This typically means: (1) pricing moves upmarket, (2) self-serve tiers receive less investment, (3) enterprise features become the focus. For solo founders and small teams, acquisition by a large security vendor is historically a signal that self-serve economics will change.
Frequently Asked Questions
Is Helicone still being actively developed?
As of May 2026, Helicone is in maintenance mode following its acquisition by Mintlify. Security updates and bug fixes continue, but new feature development has stopped. The announcement explicitly points users toward alternative platforms. Teams evaluating Helicone for production workloads should factor this in when assessing long-term reliability.
Does Portkey enforce budget caps on the $49/mo plan?
No. Budget and rate limits are listed as an Enterprise feature on Portkey's pricing page. The Production tier ($49/mo) provides observability and routing but no hard budget cap enforcement. Enforcement requires Enterprise pricing.
What is the difference between a guardrail and budget enforcement?
Guardrails in tools like Portkey validate prompt and response content — blocking PII, toxic content, or format violations. Budget enforcement checks projected spend against a remaining balance before the call fires. They are complementary features that solve different problems. A content guardrail does not stop a call that exceeds a dollar threshold.
Can I use Helicone for per-user metering?
Yes, Helicone supports cost attribution by user via custom properties on the Pro tier. You can see what each user spent in the dashboard. What you cannot do is enforce a spending limit per user — Helicone will show you the overage after the call fires, not block the call before it does.
What is the best Helicone or Portkey alternative for self-serve budget enforcement?
For teams that need hard budget caps without enterprise pricing or sales calls, noburn.dev provides pre-flight enforcement and per-user metering on a self-serve free tier. See the full LiteLLM alternatives comparison for a broader landscape view.