Vapi
Developer platform (API, SDKs, dashboard) to build, test and deploy phone and web voice AI agents; used by developers and enterprises for support, scheduling and qualification.
Score breakdown
Task success is worth up to 80 points, word error rate and latency up to 10 each.
Testing implementation
One of the calls Vapi took, end to end: the simulated caller, the agent as deployed on Vapi, the tool server it called and the grade. Callers, tools and grading are identical on every platform; only the platform and its model stack change.
Caller
C2-M1 · Conflict hit, Northgate Insurance
“My insurance company denied my claim.”
- full name
- Selma Ridgeway
- opposing party
- Northgate Insurance
- phone
- 919-555-0146
Vapi agent
Legal front desk
- STT
- Soniox stt-rt-v5
- LLM
- GPT-4.1
- TTS
- Vapi Elliot v2
- Transport
- WebSocket
- Prompt and tools
- Shared legal spec, 16 tools
Tool server
Bluejay legal systems
- Tools
- 16 HTTP tools
- State
- Seeded database per call
- Log
- Every request, keyed to the call
- Same build
- Every platform
Grade
Pass
- ✓
lookup_callerfull_name=Selma Ridgeway - ✓
check_conflictopposing_party=Northgate Insurance - ✓
escalate_to_human
Tool-call diffempty
Database-state diffmatch
Task success by industry
Share of calls where the agent completed the caller's request. Click an industry to see what its callers ask for.
Word error rate
How well the platform heard the caller: the share of the caller's words its transcript got wrong. Lower is better. Variants replay a task with background noise or a degraded line.
Latency
How fast the agent answered: time from the caller's last word to the agent's first audio, over every turn. Full marks inside the shaded band; no points in the hatched end.
Findings
Setup review
The authors' assessment of building, deploying and running the agent on Vapi: interface, developer experience, customization, going live, MCP, documentation, pricing and languages. Bluejay's voice AI engineers score it during setup from documented capabilities and hands-on use. Expand a category for what we check and how each item is scored.
UIHow well someone can build, test and review an agent in the console without writing code.8 / 10
How it is scored. Six weighted line items, each graded full, half or none: time to first agent, build without code, test in the browser, call review, errors you can see, team workspace.
| What we checked | Weight | Grade | Note |
|---|---|---|---|
| Time to first agent | 2.5 | half* 1.25 | not checked |
| Build without code | 2.0 | full* 2 | Dashboard assistant builder; Balanced preset is dashboard-only (vendor_configs/REPORT.md) |
| Test in the browser | 1.5 | full* 1.5 | Browser test call (docs) |
| Call review | 1.5 | full* 1.5 | Call logs with transcript + recording (docs) |
| Errors you can see | 1.5 | half* 0.75 | not checked |
| Team workspace | 1.0 | full* 1 | Org members (docs) |
Initial grades from the 28–29 Oct 2026 run and public docs; * = judged from docs, not yet verified on a call.
DevXWhether a developer can build, run and read back the whole agent from code.9.5 / 10
How it is scored. Six weighted line items: agent-as-code, call control, data out, API correctness, SDK quality, scoped keys and auth.
| What we checked | Weight | Grade | Note |
|---|---|---|---|
| Agent-as-code | 2.5 | full 2.5 | Assistant updated by API, read back identical, versioned (fixes/ts_date) |
| Call control | 1.5 | full* 1.5 | Create/end calls by API (docs) |
| Data out | 2.5 | full* 2.5 | Call objects with transcript (docs) |
| API correctness | 1.5 | full 1.5 | Tool schemas stored identical (vendor_configs/REPORT.md) |
| SDK quality | 1.0 | full* 1 | Python + TS SDKs (docs) |
| Scoped keys and auth | 1.0 | half* 0.5 | Public/private keys (docs) |
Initial grades from the 28–29 Oct 2026 run and public docs; * = judged from docs, not yet verified on a call.
Platform customizabilityHow far you can shape the agent: its models, conversation behaviour, prompt and tools.8 / 10
How it is scored. Seven weighted line items, including interruption handling measured on the platform's defaults during end-to-end validation.
| What we checked | Weight | Grade | Note |
|---|---|---|---|
| Model choice | 1.5 | full* 1.5 | Provider-level choice of STT/LLM/TTS (vendor_configs) |
| Conversation behaviour | 1.0 | half 0.5 | Built-in endCall hangs up immediately; escalate tools end the call after reply (fixes/vapi_endcall, vendor_configs) |
| Interruption handling | 3.0 | half* 1.5 | Not measured |
| Prompt and context | 1.0 | full 1 | Full prompt + calendar, read back identical (ts_date) |
| Custom tools | 2.5 | full 2.5 | HTTP tools, schemas intact (vendor_configs) |
| Speech controls | 0.5 | full* 0.5 | not checked |
| Agent handoff | 0.5 | full* 0.5 | Squads (docs) |
Initial grades from the 28–29 Oct 2026 run and public docs; * = judged from docs, not yet verified on a call.
Go liveEverything needed to take the agent to production and keep it running.7.2 / 10
How it is scored. Seven weighted line items, including runtime under load: the break rate at 20 concurrent calls against the 2-call baseline.
| What we checked | Weight | Grade | Note |
|---|---|---|---|
| Telephony and numbers | 2.0 | full* 2 | Numbers + SIP (docs) |
| Release management | 1.5 | full 1.5 | Each API update bumps versions v1->v2->v3 (ts_date) |
| Runtime under load | 2.5 | half 1.25 | NO_CONNECTION bursts at higher concurrency |
| Integrations at the edge | 1.0 | half* 0.5 | not checked |
| Ops and incident response | 1.5 | half* 0.75 | not checked |
| Compliance and trust | 1.0 | full* 1 | SOC 2 / HIPAA claimed (docs) |
| Support and escalation | 0.5 | half* 0.25 | not checked |
Initial grades from the 28–29 Oct 2026 run and public docs; * = judged from docs, not yet verified on a call.
MCPWhether the agent can use tools over MCP, and whether coding agents can build and run the platform through MCP.8.2 / 10
How it is scored. Four weighted line items: MCP tools on a live call, connection controls, a platform MCP server, and how much of the API that server covers.
| What we checked | Weight | Grade | Note |
|---|---|---|---|
| MCP tools on a live call | 3.5 | full* 3.5 | MCP tool (docs) |
| MCP connection controls | 1.5 | half* 0.75 | not checked |
| Platform MCP server | 3.0 | full* 3 | Official Vapi MCP server (docs.vapi.ai/sdk/mcp-server) |
| What that server can do | 2.0 | half* 1 | Coverage not measured |
Initial grades from the 28–29 Oct 2026 run and public docs; * = judged from docs, not yet verified on a call.
DocsWhether the docs are accurate, complete and usable by people and by coding agents.5.5 / 10
How it is scored. Eight weighted line items, from reference accuracy and payload schemas to change management and machine-readability.
| What we checked | Weight | Grade | Note |
|---|---|---|---|
| Reference accuracy | 2.0 | half 1 | silenceTimeout default not on API page; Balanced preset components undocumented (vendor_configs) |
| Payload and event schemas | 1.5 | half* 0.75 | not checked |
| Concepts and lifecycle | 1.0 | half* 0.5 | not checked |
| Telephony and go-live guides | 1.5 | half* 0.75 | not checked |
| Errors, limits and gotchas | 1.0 | half* 0.5 | not checked |
| Examples and quickstarts | 1.0 | half* 0.5 | not checked |
| Change management | 1.0 | half* 0.5 | not checked |
| Machine-readability | 1.0 | full* 1 | OpenAPI + llms.txt published |
Initial grades from the 28–29 Oct 2026 run and public docs; * = judged from docs, not yet verified on a call.
Pricing clarityWhether a buyer can predict what the platform will cost and see what it is costing.6.2 / 10
How it is scored. Six weighted line items, including cost per hour: our estimate from public pricing against the real bill for the scored calls.
| What we checked | Weight | Grade | Note |
|---|---|---|---|
| Public price | 2.0 | full* 2 | Public per-minute pricing |
| Every line item named | 1.5 | full* 1.5 | $0.05/min hosting + STT/LLM/TTS/telephony at cost (vapi.ai/pricing) |
| Cost per hour matches the estimate | 3.0 | half* 1.5 | Bill not reconciled |
| Usage and spend in the product | 1.5 | half* 0.75 | not checked |
| Spend controls | 1.5 | none* 0 | No cap/alerts found; prepaid credits |
| Free start | 0.5 | full* 0.5 | Free credits |
Initial grades from the 28–29 Oct 2026 run and public docs; * = judged from docs, not yet verified on a call.
Language supportWhether the agent can take a call in Spanish, French, Mandarin Chinese, Hindi or Arabic, understand it and reply in that language.7 / 10
How it is scored. One translated probe task per language, two runs each. Full when the agent understood and replied in the language, half when it understood but replied in English.
| What we checked | Weight | Grade | Note |
|---|---|---|---|
| Spanish | 2.0 | full* 2 | Graded by configuration; live probes not run |
| French | 2.0 | full* 2 | Graded by configuration; live probes not run |
| Mandarin Chinese | 2.0 | half* 1 | not checked |
| Hindi | 2.0 | half* 1 | not checked |
| Arabic | 2.0 | half* 1 | not checked |
Initial grades from the 28–29 Oct 2026 run and public docs; * = judged from docs, not yet verified on a call.
Run details
Model configuration
- STT
- Soniox stt-rt-v5
- LLM
- GPT-4.1
- TTS
- Vapi Elliot v2
- Transport
- WebSocket
Run
- Scored calls
- 630
- Runs per task
- 5
- Run dates
- 21–22 Oct 2026
- Concurrency held
- up to 14 calls
- Turns measured
- 3,647 turns