What Helicone’s acquisition means for your AI logging setup

When Helicone announced its acquisition by Mintlify on March 3, 2026, many teams relying on its AI logging proxy faced uncertainty. The transition isn’t about whether the deal is positive or negative—it’s about what it means for your workflow and whether an immediate change is required.

The reality is straightforward: Helicone’s founders have shifted their focus to Mintlify, halting new feature development for the standalone product. While security patches, bug fixes, and support for new models will continue, the roadmap has effectively paused. For teams that already have a stable setup, this might not pose an immediate problem. However, two scenarios may prompt a closer look: users of Helicone Cloud now need clarity on the product’s future direction, and self-hosted users who relied on upcoming features may feel left without options.

Interestingly, Helicone’s acquisition follows a broader trend in the observability space. Within a few months, three notable tools—Helicone, Langfuse, and Galileo—were acquired by different companies. This pattern raises an important consideration for teams evaluating replacements: prioritize solutions with transparency and self-hosting capabilities to avoid future disruptions.

Should you migrate immediately or later?

The urgency of migration depends entirely on your current setup. If you self-host Helicone and it meets your needs without requiring new features, there’s no rush. The tool will continue functioning, and you can transition at your own pace.

A stronger case for moving sooner exists if you rely on Helicone’s hosted version, depend on seamless updates for new AI providers or models, or prefer resolving the switch now rather than later. If any of these apply, the following checklist will guide your transition.

A step-by-step guide to migrating from Helicone to Spanlens

Helicone and Spanlens share a similar architecture as drop-in proxies, making the mechanical migration process relatively simple. The primary task involves updating base URLs and headers across your codebase.

1. Replace the base URL in your API client

The only mandatory change is adjusting the endpoint in your API client configurations. For example, if you were using OpenAI through Helicone, update the base URL as shown below:

// Helicone configuration (before)
const openai = new OpenAI({
  baseURL: '
  apiKey: process.env.HELICONE_API_KEY,
});

// Spanlens configuration (after)
const openai = new OpenAI({
  baseURL: '
  apiKey: process.env.SPANLENS_API_KEY,
});

2. Align your metadata headers

If your setup includes custom headers for tracking users, sessions, or prompts, map them to Spanlens’ equivalent headers:

• Helicone-User-Id → x-spanlens-user
• Helicone-Session-Id → x-spanlens-session
• Helicone-Prompt-Id → x-spanlens-prompt-version

3. Centralize provider keys in the dashboard

Unlike Helicone, which required passing both keys in each request, Spanlens simplifies key management. You can register your OpenAI or Anthropic API key directly in the Spanlens dashboard. Your application will only send the Spanlens key, reducing exposure of sensitive credentials.

4. Update endpoints for additional providers

If your workflow involves routing multiple AI providers like Anthropic, Gemini, or Azure through the proxy, ensure each provider’s base URL is updated. Spanlens provides a comprehensive migration guide with copy-paste-ready configurations for every provider.

5. Validate before decommissioning Helicone

Before removing Helicone completely, route a day’s worth of real traffic through Spanlens. Compare token and cost metrics against your provider’s invoice to confirm accuracy. Running both systems in parallel for a short period mitigates the risk of data loss or misconfiguration.

Choosing a replacement: self-hosting as a safeguard

The acquisition wave among observability tools highlights a critical lesson: data ownership matters. When evaluating replacements, prioritize tools that allow raw trace data exports without requiring support tickets and offer self-hosting options. Spanlens, for instance, is MIT-licensed and deployable with a single Docker command—a deliberate design choice to give teams control over their infrastructure.

Before finalizing your decision, consider performing a side-by-side comparison of Spanlens and Helicone. Spanlens provides a detailed breakdown covering pricing, hosting flexibility, and instrumentation capabilities. For teams ready to start, Spanlens is free to use and offers a quick setup process. Alternatively, review the migration guide to plan your transition carefully.

Ownership of your observability stack shouldn’t hinge on corporate decisions beyond your control. By selecting a tool that supports self-hosting and transparent data access, you secure your infrastructure’s future—regardless of external changes.