Hi ElevenLabs Support, I submitted a ticket (#339374) about 9 days ago regarding my AI Agent credit

[mellow mist]

AI-generated Answer for heatlab_72152

Thanks for flagging this. Here’s what the official docs say that should help while your ticket is pending:

Ticket follow‑up

• After you verify a support ticket, the Help Center notes you’ll receive a verification email and then “the team will address your inquiry as soon as possible.” If you need to follow up, use the contact form to submit another request and include your existing ticket number (#339374). (help.elevenlabs.io)

Why AI Agent credits can drop quickly

• ElevenLabs Agents (formerly Conversational AI) are billed in credits per minute of connection time (from when you click “Call AI agent” until you end the call/close the window). Typical rates per extra minute are: Creator 400, Pro 454, Scale 555, Starter 600, Free 666, Business 800. So 9 minutes can consume roughly 3,600–7,200 credits depending on plan; at the Pro rate, ~9 × 454 ≈ 4,086 credits; at Scale, ~9 × 555 ≈ 4,995. Calls made during setup/prompt testing are billed at half cost. LLM usage costs are also passed through and converted to credits. (help.elevenlabs.io)

How to verify what was charged

• In your dashboard, open Developers → Usage to see detailed Credit and Minutes Used metrics, filter by product (Agents), time range (e.g., last 15 minutes), and export CSV if needed. (help.elevenlabs.io)

Optional spend controls

• If you use Usage‑Based Billing, you can set a maximum credit threshold from Subscription → Manage Subscription to limit overage. (help.elevenlabs.io)

If you share a short snippet of your Usage Analytics (Credits + Minutes for the 9‑minute window), support can match it to your ticket more quickly when you follow up via the contact form. (help.elevenlabs.io)

Sources:
Your ticket has been verified! – ElevenLabs
How much does ElevenLabs Agents (formerly Conversational AI) cost? – ElevenLabs
How can I see information about credit usage for my account? – ElevenLabs
How do I change the usage based billing threshold for my account? – ElevenLabs
How can I see information about credit usage for my account? – ElevenLabs

[astral sierra]

Thanks for the AI-generated answer. I’ve confirmed in my Usage dashboard that my 9-minute test consumed ~4,500 credits.
Since full-length conversations are required to properly train my AI Agent, this makes it nearly impossible to prepare before launch. For 5 AI Agents, I need at least 450 minutes for each AI Agent
Could someone from the ElevenLabs team please review ticket #339374 and advise if there’s a way to:
(a) run extended tests in text-only mode without credit usage, or
(b) Receive additional training credits while the agent is in development?

[mellow mist]

AI-generated Answer for heatlab_72152

Thanks for the detailed context. I can’t access or escalate account tickets from here. To have the team review ticket #339374, please follow up via the official support form; that routes directly to the support queue. (help.elevenlabs.io)

Answers to your two questions

a) Can I run extended tests in text-only mode without using credits?

• ElevenLabs Agents supports Chat Mode (text-only) via the widget/SDK and via agent configuration. However, the documentation does not state that text-only testing is free. In fact, LLM usage costs are explicitly “passed on … converted to credits,” and setup/prompt testing calls are billed at half cost, not zero. (elevenlabs.io)
• Billing for Agents is measured by connection duration (from “Call AI agent” start until the call/window ends). This explains rapid credit use during long test sessions; to limit usage, keep connections short and end sessions when idle. (help.elevenlabs.io)
• If you want structured, text-only evaluations while iterating, use the Simulate Conversations and Agent Testing features to run scripted simulations and test suites. These are designed for development workflows; the docs do not represent them as free. (elevenlabs.io)

b) Can I receive additional credits while the agent is in development?

• Enable usage-based billing (Creator plan and above) to exceed your monthly quota with per-1,000-credit pricing; you can also set a cap. (help.elevenlabs.io)
• Consider the Startup Grants program (application-based): 12 months with 33,000,000 characters (stated as “over 680 hours of Agents”). This is intended to cover build/test/launch for startups. (elevenlabs.io)
• If you prefer a plan with more included Agent minutes immediately, the API pricing page lists included monthly minutes by tier (e.g., Scale includes 3,600 minutes; Business 13,750 minutes). (elevenlabs.io)

If you’d like, share your current plan and whether you’ve enabled usage-based billing; I can map your target 2,250+ minutes against documented credit/minute rates and suggest the most cost-effective documented path. (help.elevenlabs.io)

Sources:
How can I contact support? – ElevenLabs
Chat Mode | ElevenLabs Documentation
How much does Conversational AI cost? – ElevenLabs
Simulate Conversations | ElevenLabs Documentation
What is usage based billing? – ElevenLabs
Build AI voices into your product with our Grants program | ElevenLabs
ElevenLabs API Pricing — Build AI Audio Into Your Product
How much does Conversational AI cost? – ElevenLabs

[astral sierra]

can you explain more about the Startup Grants program (application-based): 12 months with 33,000,000 characters (stated as “over 680 hours of Agents”). This is intended to cover build/test/launch for startups.

[mellow mist]

AI-generated Answer for heatlab_72152

Here’s what’s officially published about the ElevenLabs Startup Grants program:

What you get

• Duration: 12 months. (elevenlabs.io)
• Usage: 33,000,000 characters of audio generation (described as “over 680 hours of Agents”), with scale-level benefits such as higher concurrency limits and improved support. Valued at $4,000+. (elevenlabs.io)
• Credits apply across the ElevenLabs platform for building with Conversational AI and other audio tools. (elevenlabs.io)

Eligibility

• Open to startups and small companies with fewer than 25 employees. Monetized, long-term product use cases only (no one-off projects; not for products aimed at children under 13). Existing Enterprise customers aren’t eligible. One application per company. (elevenlabs.io)

Application and timeline

• Applications are reviewed on a rolling basis; decisions are typically communicated within one week. If accepted, your account is upgraded and grant credits applied immediately. (elevenlabs.io)

After the grant / hitting limits

• If you exhaust the 33M characters early, you can talk to sales about scaling or move to a paid plan; otherwise the account converts to the Free plan at the end of the 12 months. (elevenlabs.io)

Program terms

• Recipients must display the “ElevenLabs Grants” logo in the website footer with a link to ElevenLabs for at least 12 months from approval. (elevenlabs.io)

For context on the recent expansion to “12 months” and “over 680 hours of Conversational AI audio,” see the official announcement. (elevenlabs.io)

Sources:
Build AI voices into your product with our Grants program | ElevenLabs
Build AI voices into your product with our Grants program | ElevenLabs
ElevenLabs Startup Grants: 680hrs of free AI audio | ElevenLabs
Build AI voices into your product with our Grants program | ElevenLabs
Build AI voices into your product with our Grants program | ElevenLabs
ElevenLabs Startup Grants: 680hrs of free AI audio | ElevenLabs
Build AI voices into your product with our Grants program | ElevenLabs
ElevenLabs Startup Grants: 680hrs of free AI audio | ElevenLabs

[astral sierra]

how can i get it?

[mellow mist]

AI-generated Answer for heatlab_72152

You can apply directly via ElevenLabs Startup Grants. Here’s how:

• Apply: Go to the Startup Grants page and click “Apply now,” then submit details about your product, team, and growth plan. You’ll be notified of a decision within about one week. (elevenlabs.io)
• What you get if accepted: 12 months of access with 33,000,000 characters (over 680 hours of agents) plus higher concurrency limits and improved support. Credits are applied to the ElevenLabs account you used in the application. (elevenlabs.io)
• Eligibility highlights:
  • Fewer than 25 employees at the time of receiving the grant
  • Must have a monetized/business use case and apply with a business email
  • One application per company; existing Enterprise customers aren’t eligible
  • No projects targeted at minors (under 13)
  • Rolling applications (no deadline) and worldwide availability noted in the program update (elevenlabs.io)
• After the grant: At the end of 12 months—or if you exhaust the 33M characters earlier—your plan converts to Free unless you choose a paid plan; you can also contact sales to scale. (elevenlabs.io)

Program details were expanded on Aug 12, 2025 (now 12 months/33M characters). (elevenlabs.io)

Sources:
Build AI voices into your product with our Grants program | ElevenLabs
Build AI voices into your product with our Grants program | ElevenLabs
Build AI voices into your product with our Grants program | ElevenLabs
Build AI voices into your product with our Grants program | ElevenLabs
ElevenLabs Startup Grants: 680hrs of free AI audio | ElevenLabs

[astral sierra]

witout the startup, can i get free credit for text testing

[mellow mist]

AI-generated Answer for heatlab_72152

Short answer: generally no. Outside of the Startup Grants program, ElevenLabs doesn’t offer ad‑hoc free testing credits. Here are the officially supported options:

• Free allowances you already have

  • Agents: the Free tier includes 15 minutes; calls made during setup/prompt testing are billed at half cost, but still consume minutes/credits. Billing is by connection duration and also passes through LLM costs as credits. Free/Starter tiers can’t exceed their included minutes. (help.elevenlabs.io)

• No “text‑only” or free preview mode for audio

  • There isn’t a way to preview generations without deducting quota; pressing Generate consumes credits. The website offers only two free regenerations in Speech Synthesis under specific conditions (not via API). (help.elevenlabs.io)

• Mobile app free characters (separate from web/Agents)

  • The ElevenLabs mobile app starts with 10,000 free characters/month, but ElevenReader/App credits are separate and don’t apply to the web platform or Agents usage. (help.elevenlabs.io)

• Other programs that can provide free access (if eligible)

  • Nonprofits and certain accessibility use cases can apply to the Impact Program for free licenses. (elevenlabs.io)

• Free plan/discounts

  • A free plan is available to all users; occasionally there are first‑month discounts on paid plans (check your Subscription page). (help.elevenlabs.io)

If you need account‑specific help (e.g., credit review), please contact support via the help‑center form. (help.elevenlabs.io)

Sources:
How much does ElevenLabs Agents (formerly Conversational AI) cost? – ElevenLabs
Is there a way to preview audio without losing quota before downloading? – ElevenLabs
What is the ElevenLabs App? – ElevenLabs
On a mission to help 1 million people reclaim their voices | ElevenLabs
Do you offer discounted or free plans? – ElevenLabs
How can I contact support? – ElevenLabs

[astral sierra]

As I mentioned, I trained the AI for only about 4 minutes, and it consumed more than 3,300 credits, i tried to attach an image. Can you guide me on how to train the AI agent using audio without using so many credits?

[mellow mist]

AI-generated Answer for heatlab_72152

Short answer: 3,300 credits for ~4 minutes is consistent with ElevenLabs Agents’ minute billing plus LLM passthrough. To reduce cost while iterating on an agent that you ultimately want to use with audio, use text-only (Chat Mode) for most training and reserve brief voice sessions for final checks.

Why you saw ~3.3k credits

• Agents bill by connection duration (from clicking “Call AI agent” until you end/close the window). Per-minute rates depend on plan; e.g., 400–800 credits per extra minute. LLM usage is passed through and converted to credits. So ~4 minutes can land near 1,600–3,200 credits for audio minutes alone, plus LLM credits. Calls made during setup/prompt testing are half cost. (help.elevenlabs.io)

How to train/iterate with fewer credits

• Use Chat Mode (text-only) while you iterate prompts, tools, and knowledge. Chat Mode runs without audio input/output, so you can test agent behavior without incurring voice call minutes; switch back to voice only for short validation runs. (elevenlabs.io)
• When you do need audio, run tests from the Setup & Prompt Testing environment (half-cost minutes). Keep each call short and end the session as soon as you’re done to avoid paying for idle connection time. (help.elevenlabs.io)
• Minimize silence/idle time. The platform reduces LLM inference during silence; those periods are billed at 5% of the usual per‑minute rate. Avoid leaving the session open between tests. (elevenlabs.io)
• Pick a lower‑cost LLM for iteration and keep prompts/responses concise. ElevenLabs passes LLM costs through; the LLM optimization guide covers model selection, prompt shortening, and RAG so you send/receive fewer tokens. (elevenlabs.io)
• Monitor exactly where credits go (by product, minutes, API key, voice) in Usage Analytics to confirm the driver of your spend and validate improvements. (help.elevenlabs.io)
• If you must test with audio continuously, budget using your plan’s minutes/credit rates and consider enabling usage-based billing only if needed (with thresholds) to prevent surprise overages.

[astral sierra]

I used Chat Mode (text-only), and it consumed nearly 5000 credits for about 9 minutes. I saw that you are searching the web. Can I get assistance from the company itself? mt existing ticket number (#339374), and I didn't receive any response or assistance.

[mellow mist]

AI-generated Answer for heatlab_72152

[astral sierra]

I used the Chat Mode (text-only), and it consumed nearly 5000 credits for about 9 minutes. I saw that you are searching the web. Can I get assistance from the company itself? mt existing ticket number (#339374), and I didn't receive any response or assistance.

[mellow mist]

AI-generated Answer for heatlab_72152

[astral sierra]

Can you answer my questions?

[mellow mist]

AI-generated Answer for heatlab_72152

Yes—I'll answer strictly from ElevenLabs’ official docs and link every source.

What likely drove your credit usage

• ElevenLabs Agents (formerly Conversational AI) bills live conversations by connection time, at a plan‑dependent credits/minute rate (e.g., Creator 400, Pro 454, Scale 555, Business 800). 9 minutes at Scale is about 9 × 555 ≈ 5,000 credits, which matches what you saw. LLM usage is billed separately and converted to credits. Calls during setup/prompt testing are half‑price. (help.elevenlabs.io)
• If you were chatting while a conversation connection stayed open, time‑based charges still accrue; long silences after 10 seconds drop to 5% of the normal per‑minute rate until voice activity resumes. (help.elevenlabs.io)
• General credit rules and other product-specific rates (e.g., Voice Changer/Isolator at 1,000 credits/min) are detailed here. (help.elevenlabs.io)

If you used “Chat Mode (text-only)”

• Text messages in Agents have a per‑message price (e.g., 0.3–0.4 cents, plan‑dependent). LLM costs are passed through separately. If a voice session is also connected, per‑minute audio charges can apply concurrently. (elevenlabs.io)

Ways to reduce credits while you build/test

• Use the setup/prompt‑testing flow (half cost) and end sessions as soon as you finish a test. (help.elevenlabs.io)
• Let silent periods elapse when you’re not interacting (post‑10s silence billed at 5%). (help.elevenlabs.io)
• Optimize LLM spend: pick lower‑cost models, keep prompts concise, enable RAG to pass only relevant snippets, and limit output length. (elevenlabs.io)
• Ensure your agent requires authentication so others can’t connect and consume your credits. (elevenlabs.io)
• If you routinely exceed quota, consider usage‑based billing with plan‑specific overage pricing so you control ceilings. (help.elevenlabs.io)

Getting help from ElevenLabs on your ticket

• I can’t see internal tickets, but you can contact support via the official contact form (include your existing ticket number, screenshots, timestamps, and the agent ID). You can also use the Help Center chatbot for general questions. (help.elevenlabs.io)

If you share your plan, whether a voice connection was active during those 9

[astral sierra]

Hi ElevenLabs team,

I’m testing an AI Agent and noticed very high credit usage. For example, a ~9-minute test call consumed about 5,000 credits.

From the docs I understand:

Agents bill by connection time (400–800 credits/min depending on plan).

After 10s of silence, usage drops to 5% but still accrues.

Text-only mode is per-message, but if voice is active it also adds per-minute charges.

Setup/prompt-testing is half cost, but I still need long sessions for training.

My challenge: I need to run ~450 minutes of training per agent (about 10 sessions × 45 minutes), and at these rates the cost adds up very quickly.

👉 Could you please confirm if there are any cost-reduction options or special testing credits available for long training sessions?

Ticket number: [#12345]

Thanks a lot!

[mellow mist]

AI-generated Answer for heatlab_72152

Thanks for the details. Here’s what the official docs confirm and the options available to reduce costs.

What your usage implies

• Voice calls are billed by connection duration. Minutes are measured from “Call AI agent” until the call/window ends. This matches your ~9 min ≈ ~5,000 credits if you’re on a plan with ~555 credits/min (Scale). (help.elevenlabs.io)
• During Setup & Prompt Testing, calls are billed at half cost. (help.elevenlabs.io)
• In multimodal (text + voice), text messages are billed per message; per‑minute voice pricing still applies to the connection time. Silence over 10 seconds is charged at 5% of the usual per‑minute rate until speech resumes. (elevenlabs.io)

Ways to lower training costs (documented)

• Use text‑only testing to avoid per‑minute voice billing: start sessions in Text‑only mode (conversations initiated via text do not open voice). For automated evaluation, use the Simulate Conversations endpoints (text‑based, batch/streaming). You’ll incur message/LLM costs but not voice‑minute charges. (elevenlabs.io)
• Keep connection time tight: add an “End Call” tool/prompt so the agent hangs up when a task completes, and use shorter turn‑timeouts to avoid long connected idles. (elevenlabs.io)
• Leverage half‑cost Setup & Prompt Testing for iteration sessions. (help.elevenlabs.io)
• Make use of the 5% silent‑period billing by avoiding open lines with background noise; true silence >10s is discounted automatically. (elevenlabs.io)
• Optimize LLM spend: pick lower‑cost models, trim context, and use knowledge grounding; LLM costs are passed through and converted to credits. (elevenlabs.io)
• Avoid accidental over‑charges: leave Burst Pricing off during testing (overflow calls cost 2×), and keep authentication enabled so others can’t consume your credits. (elevenlabs.io)

Plan/volume options and grants

• Lowest per‑minute rates are on Business annual ($0.08/min) with additional discounts at higher volumes and Enterprise. (elevenlabs.io)
• Startup Grants: currently advertised as 12 months with 33M free credits (≈680+ hours of agent audio) for eligible startups. Apply if you qualify. (elevenlabs.io)

Getting help on ticket #[12345]
We can’t access internal tickets here. To have the team review and follow up, please submit/append via the Help Center contact form (include your ticket ID and links/screenshots of the specific sessions). (help.elevenlabs.io)

If you share your current plan and whether those 45‑minute training runs are voice or text‑only, I can map the expected credits precisely using the per‑minute and per‑message tables from the docs.

[astral sierra]

Question:
Hello, I want to clarify the credit usage for testing my AI Agent in text-only mode (no voice connection).

How many credits per minute (or per message) does text-only testing consume?

Is billing based purely on the number of text messages exchanged with the LLM, or is there also a per-minute connection charge like in voice mode?

Could you provide an approximate credit cost for, say, a 10-minute text-only conversation with around 20 messages?

Thank you for clarifying!

[mellow mist]

AI-generated Answer for heatlab_72152

Short answer

• Text-only (Chat Mode) is billed per text message plus LLM usage. There is no per‑minute “call” charge in chat-only mode; per‑minute billing applies to voice calls. (elevenlabs.io)

Details

• Per-message platform fee:
  • Free/Starter: 0.4 cents per message
  • Creator/Pro/Scale/Business: 0.3 cents per message
  • Enterprise: custom
    These prices apply to text messages; LLM costs are passed through separately. (elevenlabs.io)
• LLM costs:
  • Billed by tokens (USD per 1M tokens) for the model you choose (OpenAI, Anthropic, Google, etc.) and converted to credits on your account. You can estimate with the “Calculate LLM Usage” endpoint or check the LLM pricing tables. (elevenlabs.io)
• Voice call billing (for comparison): charged per minute from “Call AI agent” click until call end; setup/prompt‑testing calls are half‑price; long silences drop to 5% of the usual per‑minute rate. This is voice-only behavior, not chat-only. (help.elevenlabs.io)

Approximate cost for a 10‑minute text-only session with ~20 messages

• ElevenLabs per‑message fee only (excluding LLM tokens): 20 × $0.003–$0.004 = $0.06–$0.08 total. (elevenlabs.io)
• If you want the equivalent in credits, divide by your plan’s $/1,000‑credits rate (usage-based billing):
  • Creator ($0.30/1,000 credits): $0.06–$0.08 ≈ 200–267 credits
  • Pro ($0.24/1,000 credits): ≈ 250–333 credits
  • Scale ($0.18/1,000 credits): ≈ 333–444 credits
  • Business monthly ($0.12/1,000 credits): ≈ 500–667 credits
    These conversions use the official usage‑based pricing per plan. (help.elevenlabs.io)

Notes

• Total cost in text-only mode is independent of elapsed minutes; it depends on message count plus LLM token usage.
• To estimate the LLM portion precisely for your prompt sizes/model, use the “Calculate LLM Usage” API or the LLM price tables referenced above. (elevenlabs.io)

Sources

• Chat Mode guide (text-only, no audio I/O). (elevenlabs.io)
• Agents Platform overview (per-message pricing; silence discount for voice). (elevenlabs.io)
• “How much does Agents cost?” (per‑minute call billing rules; LLM costs converted to credits). (help.elevenlabs.io)
• LLM pricing and estimation. (elevenlabs.io)
• Usage‑based

[lofty cloud]

Hey Team, is simulate_conversation api work with the alpha workflow ?

[mellow mist]

AI-generated Answer for bright_moose_26625

Short answer: The docs don’t explicitly confirm that simulate_conversation runs an agent’s (alpha) visual Workflow graph end‑to‑end. What is documented is that Simulate Conversation executes a text-based interaction against an agent and returns the full transcript plus tool_calls/results, which you can use to verify behaviors like transfers and tool dispatch. For structured workflow validation, the Agent Testing framework (scenario tests and tool‑call tests) is the documented path.

References:

• Simulate Conversation API and guide (returns transcript, tool_calls/results). (elevenlabs.io)
• Agent Workflows overview (visual graph with Subagent, Dispatch Tool, Agent Transfer, Transfer to number nodes). (elevenlabs.io)
• Agent Testing (scenario and tool‑call testing; mentions verifying tools such as transfer_to_number). (elevenlabs.io)
• Agent‑to‑agent transfer workflow. (elevenlabs.io)

If you need an officially guaranteed workflow-by-workflow simulation entry point, it isn’t stated in the docs today.

Sources:
Simulate conversation | ElevenLabs Documentation
Agent Workflows | ElevenLabs Documentation
Agent Testing | ElevenLabs Documentation
Agent-agent transfer | ElevenLabs Documentation

[lofty cloud]

Hi team. I tested 50 times ! How can we dirce a simulation to test a mcp tools ? Is it possible ?