Comparison
Verum vs Poe: a Poe alternative that uses your own API keys
Both let you reach many AI models from one place. The difference is who holds the keys and who you pay for tokens. Poe bundles model access into its own platform. Verum connects to the labs with your API keys and charges a flat $12/month for the workbench.
Verum vs Poe: feature comparison
| Capability | Verum | Poe |
|---|---|---|
| Bring your own API keys (BYOK) | Yes — your keys, encrypted at rest | No — access is bundled into the platform |
| Who you pay for tokens | The labs directly, at their list price | The platform, via its own credit system |
| Per-call cost + token receipt | Yes — cost, tokens and latency per stage | Credits are deducted per message |
| Blind side-by-side comparison + vote | Yes — Truth Arena, names hidden until you vote | Not a blind, vote-recorded comparison |
| Personalised model recommendation | Yes — Best pick, learned from your own votes | Model choice is manual |
| Multi-model pipelines | Yes — panel, merge, judge, research → mind map | Single-bot conversations |
| Breadth of ready-made bots | Smaller — Verum is a workbench, not a bot marketplace | Large public bot directory |
| Mobile apps | Responsive web app | Native mobile apps |
Verum column reflects shipped behaviour as of 24 July 2026. Poe changes frequently — check poe.com for current details. Poe is a trademark of Quora, Inc. Verum is not affiliated with or endorsed by Quora.
BYOK vs platform credits: which costs less?
It depends on how much you use. A bundled subscription is predictable and is usually cheaper for light, occasional use — you pay one price and stop thinking about it. BYOK is better when you run a lot of tokens or want to see exactly what each answer cost, because you pay the lab's list price with no intermediary and every call is itemised.
Verum shows cost, tokens and latency for every stage of every run, and the analytics dashboard breaks spend down by model and by feature. If you have never wanted to know what a single answer cost, you probably do not need this.
Blind model comparison, with your votes
Truth Arena runs two models on the same prompt with identical system prompts and settings, and hides which is which. You vote, then the names are revealed. Those votes accumulate into a personal ranking — Best pick — so recommendations reflect what you actually preferred rather than a public leaderboard.
When Poe is the better choice
Verum is not the right tool for everyone. Choose Poe if you want native mobile apps, a large directory of ready-made bots, or a single subscription without managing API keys at all. Verum assumes you are willing to create keys at each lab, and rewards that with direct pricing, receipts and comparison tooling.
Frequently asked questions
What is the best Poe alternative for people who already have API keys?
Verum is built for exactly that case. You connect your own OpenAI, Anthropic, Google, xAI, Groq or Perplexity keys and pay those labs directly for tokens, while Verum charges a flat $12/month for the interface. If you do not already have API keys, a bundled platform like Poe is usually simpler to start with.
Does Verum resell model access or add a markup on tokens?
No. Calls are made with your keys against your own lab accounts, so you pay the lab's list price. Verum's subscription covers the workbench itself.
Where are my API keys stored?
Keys are encrypted at rest with AES-256-GCM and stored server-side, because Verum proxies provider requests on your behalf. They are never shown back to you in full — only a masked preview — and each account's data is isolated at the row level.
What is Truth Arena?
Two models answer the same prompt side by side with their names hidden. You vote for the better answer, and only then are the names revealed. Those votes build a personal ranking that powers Best pick.
Can I see what each answer cost?
Yes. Every run records input and output tokens, cost, latency and which key was used, per stage. You can export the receipt and review spend by model or feature in the analytics dashboard.
Use your own keys. See what every answer costs.
$12/month for the workbench. Tokens are billed by the labs, at their price.
Create a free account