GTM Hub
What is Simli?
Simli is a real-time speech-to-video API that renders lip-synced 3D Gaussian-splat avatars from audio with ~150 ms latency. Developers pair it with their own STT, LLM, and TTS stack to add a visual face to voice agents—ideal for GTM teams building low-latency sales bots, coaching apps, and support assistants in LiveKit or Pipecat pipelines.
Key features
Low-Latency Speech-to-Video
Send audio frames and receive avatar video in about 150 ms, minimizing perceptible delay atop existing voice agent latency.
Bring-Your-Own-Agent Stack
Simli handles face rendering only; you keep preferred LLM, TTS, and orchestration—reducing vendor lock-in vs full-stack avatar agents.
Official Integrations
First-party guides and SDKs for LiveKit, Pipecat, JavaScript, Python, OpenAI, and ElevenLabs reference architectures.
3D Gaussian Avatar Pipeline
Builds consistent 3D avatars from a single photo for stable angles across portrait, landscape, and square streams (e.g., 1366×768).
Custom Avatars
Paid tiers include monthly custom avatar generation quotas with concurrent generation limits scaling by plan.
Concurrency Controls
Plans scale concurrent sessions from 1 on Free to 50 on Scale—match plan to peak simultaneous website visitors.
Enterprise Privacy Options
Enterprise offers zero data retention, 99.95% uptime SLA, and pricing down to ~$0.005/min at volume.
Simli pricing
Simli prices real-time streaming in included monthly minutes with per-minute overage semantics on paid tiers. Free: $0, 50 min/mo, 1 concurrent session, template avatars. Hobby: $10/mo, 1,000 min (~$0.01/min), 2 concurrent sessions, 1 custom avatar/mo. Pro: $49/mo, 5,500 min (~$0.0095/min), 10 concurrent sessions, 5 custom avatars/mo. Scale: $249/mo, 27,500 min (~$0.009/min), 50 concurrent sessions. Enterprise: custom, from ~$0.005/min at scale with higher concurrency and SLA. Billing may enforce minimum session increments (e.g., 1-minute minimums cited in industry comparisons).
| Plan | Price | What's included |
|---|---|---|
| Free | $0/month | 50 streaming minutes/mo, 1 concurrent session, template avatars, personal/evaluation use |
| Hobby | $10/month | 1,000 minutes/mo, $0.01/min effective, 2 concurrent sessions, 1 custom avatar/mo |
| Pro | $49/month | 5,500 minutes/mo, $0.0095/min, 10 concurrent sessions, 5 custom avatars/mo |
| Scale | $249/month | 27,500 minutes/mo, $0.009/min, 50 concurrent sessions, 25 custom avatars/mo, 5 concurrent avatar generations |
| Enterprise | Custom | As low as $0.005/min, high concurrency, zero data retention option, 99.95% SLA, dedicated support |
Pros & cons
Pros
- + Among the lowest per-minute real-time avatar rates when plans are fully utilized.
- + Architecture fits modern voice-agent stacks (LiveKit/Pipecat) without replacing LLM choices.
- + Sub-200 ms rendering latency keeps conversational UX credible on customer-facing sites.
- + Generous free 50 minutes/mo for engineering proof-of-concepts.
- + Custom avatar quotas on every paid tier support branded agent faces.
- + Enterprise zero-retention narrative helps regulated industries pilot visual agents.
Cons
- − Not a full video studio—no async script-to-campaign editor for marketing batch jobs.
- − You must build and host STT/LLM/TTS; Simli is a face layer only.
- − Concurrent session caps punish viral traffic unless Enterprise negotiated.
- − Included minutes waste if production traffic is spiky month-to-month.
- − Privacy policy notes user content may train models—review before regulated deployments.
- − Less turnkey than D-ID or HeyGen for non-technical marketing teams.
Who is Simli best for?
- • Engineering-led GTM teams shipping voice agents that need a face on the website.
- • Edtech and coaching startups prototyping mock interviews with Pipecat.
- • Customer support innovators adding visual presence to existing LLM chatbots.
- • Startups optimizing COGS on per-minute avatar spend at Scale tier volume.
- • Agencies integrating real-time avatars into client apps via REST/WebRTC SDKs.
- • Builders already standardized on LiveKit rooms who want drop-in visual output.
Frequently asked questions
Does Simli include an LLM or voice?
No. Simli renders avatar video from audio you send. You supply speech recognition, language model, and text-to-speech—or use partner integration guides.
What resolution and formats are supported?
Streams commonly run at 1366×768 16:9 with portrait and square canvases available. Latency targets ~150 ms from audio in to video out.
How is usage billed?
Plans include a pool of monthly streaming minutes priced effectively at $0.01 down to $0.005/min. Check docs for minimum billing increments per session.
Can I create custom avatars?
Hobby and above include monthly custom avatar generation slots; Scale increases concurrent avatar generation for faster iteration.
