Overview
Chert and Humalike are two emerging products that aim to enhance AI agents, but they tackle very different problems. Chert focuses on giving AI agents eyes by enabling them to make and receive FaceTime calls, allowing them to see the caller's environment and provide visual assistance. Humalike, on the other hand, focuses on giving AI agents social skills, making them feel more human in conversations by understanding social norms, remembering interactions, and reading social cues.
Feature Comparison
| Feature | Chert | Humalike |
|---|---|---|
| Core Focus | Video AI agents for FaceTime calls | Behavioral infrastructure for humanlike AI agents |
| Primary Interface | FaceTime video calls with camera context | APIs and models that integrate into any agent stack |
| Deployment | Provisioned FaceTime lines with managed control plane | APIs that are model-agnostic and stack-agnostic, with one-shot integration for Hermes |
| Key Capabilities | Visual understanding, remote support, field service, telehealth intake, guided onboarding | Norms, persona, theory of mind, social observability, social memory, social signals |
| Target Use Cases | Remote support, field service, telehealth, onboarding, visual inspections, customer success | Gaming characters, AI coworkers, therapy/care, companions, ed-tech, humanoids, creators/communities |
| Availability | Controlled private preview; not generally available yet | Behavioral APIs coming soon; free credits to start building |
| Compliance | Not specified | SOC 2 Type II and ISO 27001 in progress |
Pricing
Chert: Chert is in a controlled private preview, and pricing details are not publicly disclosed. Interested users can book a demo or try the browser preview to test the agent without placing a real call.
Humalike: Humalike offers $20 in free credits for new accounts, no card required. This allows developers to wire the APIs into their agents and ship something real. After that, usage-based pricing for the behavioral APIs is expected, but specific rates are not yet published.
Pros and Cons
Chert
Pros:
- Enables visual context for AI agents, solving problems that are hard to describe verbally.
- Simple three-step setup: configure prompt, choose persona, deploy with a line.
- Supports both inbound and outbound FaceTime calls in controlled workflows.
- Browser preview allows testing without placing a real call.
Cons:
- Not generally available; limited to a controlled private preview.
- Limited to FaceTime platform; not cross-platform.
- Live execution requires explicit authorization; accepting mode is not enabled by default.
Humalike
Pros:
- Model-agnostic and stack-agnostic APIs that can be composed into any agent.
- Focus on social behaviors like norms, theory of mind, and social memory, making agents feel more human.
- One-shot integration with Hermes for quick deployment.
- Free credits to start building without upfront cost.
Cons:
- Behavioral APIs are 'coming soon' and not yet fully available.
- Compliance certifications are in progress, not yet complete.
- Requires integration effort to add behavioral layer to existing agents.
Verdict
Choose Chert if your priority is enabling AI agents to see and interact via video calls, especially for visual support tasks. Choose Humalike if you need to enhance your agents' social intelligence and human-like behavior across various platforms. Both are early-stage but address different gaps in the AI agent landscape.