4 Portkey Alternatives When You Actually Wanted Observability (2026)

Portkey is the strongest LLM gateway in this space. It routes to 1,600+ models through one API, with fallbacks, caching, budgets and 50+ guardrails, and the open-source gateway is Apache-2.0 and self-hosts free. If your problem is “we call five providers, costs are exploding, and we need fallbacks and budgets in one place,” it is aimed squarely at you.

The reason people look for alternatives is a mismatch between what they expected and what self-host gives them. The open-source gateway gets you routing and a basic dashboard - the real logging, traces, analytics and retention live on the managed paid tier. Self-host Portkey expecting a full observability platform and you get a proxy and a basic dashboard, then discover the analytics you wanted are a paid feature at $49/mo. There is a second sharp edge in the billing too - the meter caps logs, not requests, so past the cap your traffic keeps routing but your observability quietly goes dark.

That is not dishonest. It is open-core, and the free gateway is genuinely useful. But if the job you came for was tracing and logging, not routing, these are the tools built for that job. I have matched each to why you are here.

The short version

Tool	Best for	Starting price	Self-host observability	License
Portkey	Multi-provider gateway with budgets	Free / $49/mo	No - gateway only	Apache-2.0 gateway
Langfuse	Full self-hosted observability, free	Free / $29/mo	Yes, near-complete, MIT	MIT
Opik	Cheapest managed cloud, full OSS	Free / $19/mo	Yes, full features, no gates	Apache-2.0
Helicone	Nobody new - maintenance mode	Free / $79/mo	Yes, but frozen	Apache-2.0

If you wanted free self-hosted observability: Langfuse

This is the direct fix for Portkey’s gap. Where Portkey self-host gives you the pipe, Langfuse self-host gives you the whole product. Only three features are enterprise-gated in the self-host build - tracing, evals, prompt management, human annotation and RBAC are all free under MIT. That is the entire logs-and-traces dashboard Portkey holds back for its paid tier, running on your own infrastructure for free. On the managed side it runs about $101/mo at 1M events, cheaper than most of the category at scale, and it is framework-agnostic.

The catch is operational, not commercial. Langfuse v3 needs Postgres plus ClickHouse, Redis and S3-compatible storage - four services - and the ClickHouse migration is where self-hosters get stuck, with GitHub threads reporting memory errors and timeouts. It is also a ClickHouse subsidiary now after the January 2026 acquisition, which I would weigh on a multi-year bet. If you cannot spare the ops time, the $29/mo Core tier sidesteps the whole thing. The Langfuse page has the detail.

If you want the cheapest managed cloud: Opik

If the appeal is dropping the gateway complexity and just getting logs, Opik is the cheapest managed route. It is Comet’s open-source observability and eval platform, and its Pro cloud tier is $19/mo for 100k spans - the cheapest paid cloud of the major eval platforms, undercutting Portkey’s $49/mo Production tier. Overage is $5 per additional 100k spans. The OSS build is Apache-2.0, the most permissive license here, and Comet says the self-hosted version ships the full feature set with no gates - so unlike Portkey, self-host gets you the real logging too. It is the fastest-growing project of its peers at roughly 20.8k GitHub stars.

The honest gotcha is per-seat pricing at scale. The recurring complaint is that seat costs climb as your team grows, so the $19 headline is the small-team configuration. A few users report UI slowdown on very large projects. Model the seat cost first if you are a big team.

If you specifically want a proxy logger: know Helicone is frozen

Portkey and Helicone both put themselves in the request path, so if the proxy model is what you liked, Helicone is the obvious name to consider - and the one I am including to warn you off for new builds. It was a clean, open-source, proxy-based observability tool. Then Mintlify acquired it on 3 March 2026 and put it in maintenance mode - security patches, bug fixes and new-model support only, no roadmap, and Mintlify is actively helping customers migrate off.

For a new buyer that single fact overrides everything. The Apache-2.0 self-host still works, but self-hosting a frozen codebase whose own owner is pointing people to the exit is a weak bet. The proxy model was also a structural single point of failure - if Helicone is down, your LLM calls fail even when the provider is healthy, and every proxied call adds latency (Helicone cites around 10ms). If you are on it, use the runway to plan a migration. If you are evaluating fresh, skip it.

So which one?

• You need a multi-provider gateway more than a tracer - stay on Portkey, and expect to pay for the real observability on the Production tier. It is the best gateway in this set.
• You want the full logs-and-traces product self-hosted for free - Langfuse, if you can run the four-service stack.
• You want managed logging for the least money - Opik at $19/mo, cheaper than Portkey’s paid tier and with a clean Apache-2.0 self-host as your exit.
• You want a proxy logger - do not start on Helicone; it is frozen. Go to Langfuse or Opik instead.

One pairing worth naming, because it is often the real answer - keep Portkey as the gateway for routing and budgets, and feed a dedicated observability backend like Langfuse over OpenTelemetry. Portkey’s OTel support enriches traces with cost and token metrics per the GenAI conventions, so the handoff is clean. Every price and date here is from each vendor’s public pages, verified on 26 July 2026. This category ships breaking changes monthly - we re-verify every 30 days.