Helicone Alternatives
Helicone went into maintenance mode after Mintlify's acquisition. Here's where five real alternatives stand on price.
Mintlify acquired Helicone on March 3, 2026, and Helicone's own announcement says the product now ships security updates and bug fixes only, with no new development, while the team works with each customer on a migration path. Most of the Helicone alternatives roundups still ranking for this search were written before that changed, or leave it out entirely. This page verifies Langfuse, LangSmith, Portkey, Arize AI, and Braintrust against their own current pricing pages, and covers the one gap none of them, or Helicone before it, ever solved: what your own paying customer gets to see.
Why teams are looking now
What actually happened to Helicone, and why the timing matters
Helicone works as a lightweight proxy in front of a team's model calls, adding request logging and cost tracking with a small integration change, usually just a base URL swap, rather than a full SDK rewrite. Its free Hobby tier covers ten thousand requests, one gigabyte of storage, and seven days of retention on a single seat, and its Pro tier runs $79 a month for unlimited seats and thirty-day retention. For a lot of teams, that combination, a fast integration and clear request-level cost tracking, was and still is a perfectly reasonable choice, and there was no particular reason to look elsewhere.
That changed on March 3, 2026, when Mintlify acquired Helicone. Mintlify's own announcement is specific about what the acquisition means in practice: Helicone moves into maintenance mode, still receiving security updates and bug fixes, but with no new feature development going forward, and Mintlify says it will work directly with existing customers on a migration path to another platform. That is the acquiring company's own stated plan, not a guess pieced together from a quiet changelog or a Reddit thread. Helicone's founders joined Mintlify to build something else entirely, which is a fair signal that Helicone itself was not the product either side saw a long runway for. A team that chose Helicone specifically for a feature still on its roadmap now has a vendor-confirmed reason to plan the move, which is the context most of the roundups still ranking for "Helicone alternatives" either skip or only gesture at.
Five alternatives, verified against their own pricing pages
What Langfuse, LangSmith, Portkey, Arize AI, and Braintrust actually cost
Every figure below comes from each vendor's own pricing page as of this page's publish date, not a secondhand number a roundup forgot to refresh.
Langfuse
Tracing and evaluation in one open source tool, with a genuinely free self-hosted option and no usage cap. Cloud pricing starts at $0 for fifty thousand units a month across two seats with thirty days of data access, Core runs $29 a month for a hundred thousand units and unlimited seats with ninety days of access, Pro is $199 a month for the same hundred-thousand-unit allowance with three years of access, and Enterprise starts at $2,499 a month.
LangSmith
Built by the LangChain team, so it reads LangChain and LangGraph internals natively rather than through a generic hook. Developer is free for one seat and five thousand base traces a month, Plus runs $39 a month per seat with unlimited seats and ten thousand base traces included, and Enterprise is custom priced with self-hosted and hybrid deployment on top.
Portkey
An AI gateway built around routing and caching first, request logging second, which makes it the closest architectural match to Helicone's own proxy. The open source version self-hosts free with no limit, Developer is free forever for ten thousand recorded logs a month, Production runs $49 a month for a hundred thousand logs with $9 per additional hundred thousand past that, and Enterprise starts around ten million logs a month, custom priced.
Arize AI
Unlimited seats and unlimited evaluations on every tier, so cost scales with data volume instead of headcount. AX Free covers twenty five thousand spans and one gigabyte of ingestion with fifteen days of retention, AX Pro is $50 a month for fifty thousand spans and ten gigabytes of ingestion with thirty days of retention, and Enterprise is custom priced.
Braintrust
Evaluation-first, with unlimited users, projects, and experiments even on the free Starter tier, which includes ten dollars of model credit and fourteen days of retention. Pro is $249 a month flat with no per-seat charge, covering five gigabytes of processed data and fifty thousand scores before overage, and Enterprise has no published price.
What none of the six answer
Every one of them scores your agent. None of them show your customer what it's worth
Every published rundown of Helicone alternatives describes the same audience: an engineering team choosing where to send request logs and traces after Helicone stops adding new features. Latitude's rundown of six tools, TrueFoundry's rundown of five, and PostHog's own comparison all build the entire piece around integration effort, framework fit, and pricing tiers, and not one of them raises the question of whether a customer outside the company should ever see any of it. That holds across Langfuse, LangSmith, Portkey, Arize AI, and Braintrust too, and it held for Helicone itself before the acquisition: every dashboard is built for the team that shipped the agent, scoped to one account, with no notion of a paying customer as a separate audience with numbers of their own.
If you sell an AI agent inside a product other companies pay for, that gap becomes the second problem right after the migration itself gets handled. A customer renewing a contract wants their own deflection rate and cost per resolution, scoped to their own traffic only, and none of the five tools above, or Helicone, were built with a second tenant layer in mind. AiAgRe's Node SDK connects to a LangChain, LlamaIndex, or CrewAI agent the same way a proxy or a tracer does, tags each event with an organization identity and a customer identity at the point of ingestion, and turns that into white-labeled dashboard components a customer sees inside your own product instead of a shared login to whichever tool replaced Helicone. It doesn't replace the request logging and evaluation work these tools do well: most teams keep one of them running for the team that built the agent, and add AiAgRe for the customer using it.
The honest answer, not the sales pitch
Three signs the migration off Helicone isn't urgent yet
None of these show up in a feature comparison table. They show up once the migration is actually on the calendar.
Usage is light and nothing is broken
Maintenance mode still means security patches and bug fixes ship, and a team well under the free tier's ten-thousand-request allowance, with no open incidents, has real runway before this becomes urgent. Migrating on a deadline that doesn't exist yet just adds risk for no reason.
The one-line proxy integration is doing exactly what was needed
Helicone's base-URL swap was chosen for a reason: minimal integration work for request cost tracking, nothing more. Portkey is the only alternative on this page built the same way; the other four ask for a heavier SDK-based instrumentation change that isn't worth making until a specific feature gap actually shows up.
Request cost tracking was always the whole job
A team that never needed evaluation datasets, regression testing, or framework-native LangChain internals doesn't gain much by moving to a heavier tool built around those features. If cost and request logging is genuinely the entire requirement, Helicone still does that job today.
FAQs
Helicone alternatives: frequently asked questions
Common questions from teams migrating off Helicone after the Mintlify acquisition, or pricing a second tool on top of whichever one they pick.
What actually happened to Helicone, and why are teams searching for alternatives now?
Mintlify acquired Helicone, announced March 3, 2026. Mintlify's own post is specific about what changes: Helicone moves into maintenance mode, still shipping security updates and bug fixes, but with no new feature development, and Mintlify says it will work directly with each existing customer on a migration path to another platform. That is the acquiring company's own stated plan, not a rumor pieced together from a changelog going quiet. Helicone's founders, Justin Torre and Cole Gottdank, joined Mintlify to work on a different product, which is a reasonable signal that Helicone was not the long-term roadmap either side had in mind.
Is Helicone shutting down? Can I still use it today?
No, it is not shut down, and existing workspaces keep running. Maintenance mode means the request logging, cost tracking, and proxy still work, and security issues still get patched, but nothing new is being built on top of it. For a team with light usage and no incidents, there is no fire to put out this week. For a team that picked Helicone specifically because a new feature was on its roadmap, that feature is not coming, and the honest planning move is to treat this as a slow migration window rather than an emergency one.
Which Helicone alternative is the closest architectural match?
Portkey is the closest match, by a clear margin. Helicone works as a proxy sitting in front of the model calls, adding logging and cost tracking without asking for much integration work beyond swapping a base URL. Portkey is built the same way, as a gateway that adds routing and caching on top of that same proxy pattern, which makes it the option closest to a like-for-like swap. Langfuse, LangSmith, Arize AI, and Braintrust all lean more toward an SDK you instrument directly, which is a bigger integration change even though none of the four are difficult to add.
Is Langfuse, LangSmith, Arize AI, or Braintrust cheaper than Helicone?
At the free tier, Langfuse's fifty-thousand-unit Hobby plan and Arize's twenty-five-thousand-span AX Free tier both start ahead of Helicone's ten-thousand-request Hobby plan, and LangSmith's five-thousand-trace Developer tier starts behind it. Past the first paid tier the four stop lining up cleanly against Helicone's $79-a-month Pro plan: Langfuse's Core plan is $29 a month, Portkey's Production tier is $49 a month, Arize's AX Pro is $50 a month, and Braintrust's Pro plan is a flat $249 a month with no per-seat charge. None of the five meter the same unit Helicone does, a raw request through its proxy, so a same-tier dollar comparison alone undercounts what each one is actually billing for.
Does switching away from Helicone mean losing LangChain, LlamaIndex, or CrewAI support?
No, it does not: Langfuse, LangSmith, Portkey, Arize AI, and Braintrust all integrate with LangChain, LlamaIndex, and CrewAI through an SDK hook or a proxy swap, the same general pattern Helicone uses today. Moving off Helicone's proxy does not mean rebuilding an agent that already runs on any of those frameworks, and none of the five require a different tracing approach than the one already in place. AiAgRe's own Node SDK ships the same three integrations, which is why it sits alongside whichever tracer replaces Helicone rather than asking a team to change how the agent is built and observed in the first place.
Do any of these Helicone alternatives let a customer see their own dashboard?
No, none of them do: Langfuse, LangSmith, Portkey, Arize AI, and Braintrust all render results inside a workspace scoped to your own account, built for your own engineers, with no separate view for a customer outside your company, the same limit Helicone has today and had before the acquisition. AiAgRe covers that second job specifically: a Node SDK that reads the same kind of trace and cost data these tools already produce and turns it into a white-labeled dashboard your own customer sees inside your product, scoped so one customer never sees another customer's numbers.
Can I run AiAgRe alongside whichever Helicone alternative I pick?
Yes, and that is the usual setup rather than a workaround. AiAgRe's SDK taps the same underlying agent events a proxy or a tracer already reads, tags each one with an organization and customer identity at ingestion, and does not require removing whichever Helicone alternative is already in place. Most teams keep one tool running for internal debugging and cost tracking, and add AiAgRe's dashboard components for what their own customers see once the agent is live.
Does AiAgRe replace Helicone or any of its alternatives?
No, and it is not trying to. Helicone, and every alternative on this page, is built for logging, tracing, or evaluating an agent for the team that shipped it: request costs, latency, dataset scores, prompt iteration. AiAgRe starts only once that part is covered, reading the same events a tool like Portkey or Langfuse already logs and turning them into the deflection rate, cost per resolution, and resolution rate your own paying customer sees. Most teams run one of the tools above for the first job and add AiAgRe for the second, rather than asking one tool to do both.
Migrating off Helicone? Now show your customer what the agent did.
Request access and we'll walk through how AiAgRe's embed tokens map onto the trace data whichever tool you pick already produces.