Step 1 — Gather context
If references/learnings.md exists, read it first for accumulated platform knowledge.
-
Which Krisp product are you using (or evaluating)?
- A) AI Meeting Assistant (noise cancellation + transcription + AI notes)
- B) Call Center AI (accent conversion, agent assist, speech analytics)
- C) Voice AI SDK (embedding noise cancellation / accent conversion in your app)
- D) Not sure — help me pick
-
What's your goal?
- A) Set up or configure Krisp
- B) Integrate Krisp with other tools (CRM, Slack, webhooks)
- C) Troubleshoot a problem (crashes, transcription issues, missing recordings)
- D) Compare Krisp plans or evaluate pricing
- E) Build with the Krisp SDK
- What plan are you on?
- A) Free trial / evaluating
- B) Core ($8/mo annual)
- C) Advanced ($15/mo annual)
- D) Enterprise
- E) Call Center or SDK
Skip-ahead rule: if the user's prompt already contains enough context, skip to Step 2.
Step 2 — Route or answer directly
| Problem domain | Route to |
|---|---|
| Choosing between AI note-takers (Krisp vs Fireflies vs Fathom etc.) | /sales-note-taker {user's question} |
| Building a transcript pipeline to CRM or data warehouse | /sales-note-taker {user's question} |
| Reviewing a specific call for coaching | /sales-call-review {user's question} |
| General CRM-to-tool integration patterns | /sales-integration {user's question} |
| Sales coaching program | /sales-coaching {user's question} |
Otherwise, answer directly below.
Step 3 — Krisp platform reference
Read references/platform-guide.md for the full platform reference — modules, pricing tiers, integrations, data model, workflows, SDK details.
If the question involves the webhook API or Voice AI SDK, also read references/krisp-api-reference.md.
Answer the user's question using only the relevant section. Don't dump the full reference.
Step 4 — Actionable guidance
You no longer need the platform guide — focus on the user's specific situation.
For setup questions: Walk through the specific workflow step by step. Krisp installs as a virtual audio device — the main gotcha is selecting "Krisp Microphone" and "Krisp Speaker" in your conferencing app's audio settings.
For integration questions: Webhooks are the primary programmatic integration path — the current pricing page shows webhook capability starting at Core, with Advanced adding team-level webhooks (an older announcement framed the Webhook API as Business/Advanced-only; verify in-app). Krisp fires events when transcripts, notes, or outlines are generated. For CRM sync, HubSpot is available on Core+, Salesforce on Advanced+. Krisp also offers MCP integration (Claude/ChatGPT/Cursor) on Core+.
For SDK questions: Point to sdk-docs.krisp.ai. The JS SDK (WASM) is the most documented. VIVA SDK is for voice AI agents (server-side); there's also a self-serve Voice Translation API (61 languages, 60 min free credit). Auth is an API key from the developers.krisp.ai dashboard, exchanged at runtime for a short-lived session key.
For troubleshooting: Check references/platform-guide.md Troubleshooting section first, then apply the fix to the user's specific situation.
If you discover a gotcha, workaround, or tip not covered in references/learnings.md, append it there.
Gotchas
Best-effort from research — review these, especially items about plan-gated features and integration gotchas that may be outdated.
- Free tier was removed. Krisp used to offer 60 min/day free noise cancellation. Now it's a 7-day trial only — Core plan ($8/mo annual) is the entry point.
- Salesforce requires Advanced plan ($15/mo). Core only gets HubSpot, Slack, Zapier.
- Accent conversion officially covers Indian, Filipino, African, Latin American English. Both speaker-side and listener-side (real-time, launched March 2026) are offered; more accents roll out over time.
- Transcription now covers up to 95 languages via Multilingual Meeting Transcriptions (in beta as of 2026-06). English is on-device; other languages are server-side.
- Webhook capability: per the current pricing page, webhooks start at Core ("Unlimited Integrations & webhook"), with Advanced adding team-level webhooks. (An earlier Krisp announcement framed the Webhook API as a "Business"/Advanced feature — sources conflict; verify the gate in-app.)
- App can slow down laptops — Krisp uses significant CPU for real-time neural network processing. Close other heavy apps during calls.
- Noise cancellation only works with USB headsets — Bluetooth and 3.5mm jack headphones may not be supported.
Related skills
/sales-note-taker— Compare AI note-takers across vendors, plan transcript API pipelines/sales-call-review— Review a specific call for coaching/sales-coaching— Build coaching programs/sales-integration— General CRM/tool integration patterns (Zapier, webhooks, iPaaS)/sales-do— Not sure which skill to use? The router matches any sales objective to the right skill. Install:npx skills add sales-skills/sales --skill sales-do
Examples
Example 1: Setting up Krisp for a sales team
User says: "My reps work from home and clients complain about background noise on calls. How do I roll out Krisp?" Skill does: Recommends Core plan for the team ($8/user/mo annual), walks through deployment — install desktop app, select Krisp virtual audio device in Zoom/Teams/Meet settings, enable AI meeting transcription, set up HubSpot integration to auto-push meeting notes.
Example 2: SDK integration for a voice AI agent
User says: "I'm building a voice AI agent and need noise cancellation in the browser. Can Krisp help?" Skill does: Points to the JS SDK (WASM) at sdk-docs.krisp.ai. Explains the VIVA model for voice isolation + turn-taking detection. Notes 12MB package size, ~100MB memory footprint, 1.5-2ms per frame latency. Walks through createNoiseFilter API and WebRTC integration.
Example 3: Troubleshooting missing recordings
User says: "Krisp recorded 9 meetings but only 3 have transcriptions. Where did the others go?" Skill does: Checks common causes — Krisp microphone not selected as input (recordings captured silence), meetings under minimum duration threshold, storage limit hit (Core = 10 GB, Advanced = 60 GB). Recommends checking Settings → Recordings for failed entries, verifying audio input device in each conferencing app.
Troubleshooting
Transcription is inaccurate or unreliable
Symptom: AI notes miss key points, words are wrong, speaker names are mixed up Cause: Krisp transcription accuracy depends on audio quality and speaker clarity. Non-English languages use server-side processing which can be less accurate. Speaker identification degrades with 5+ participants. Solution: Ensure Krisp noise cancellation is active (improves transcription accuracy by up to 39%). Add custom vocabulary for technical terms (Settings → Custom Vocabulary, up to 750 terms). For multi-speaker meetings, ask participants to identify themselves early. Check that the correct microphone is selected.
App crashes or slows down laptop
Symptom: Krisp freezes, high CPU usage, laptop fans spinning during calls Cause: Real-time neural network processing is CPU-intensive. Conflicts with other audio processing apps (Discord voice, OBS, other noise cancellation tools). Solution: Close competing audio apps. Check for Krisp updates (Settings → About). If on an older machine, disable video recording and use transcript-only mode. Krisp's help center documents known conflicts with specific apps — check External Applications Interferences article.
Calendar records everything including non-meetings
Symptom: Krisp tries to record calendar tasks, focus time blocks, and reminders — not just actual meetings Cause: Krisp auto-records based on calendar events without distinguishing meeting types Solution: Use manual recording mode instead of auto-record. Or filter calendar events by excluding events without video links (Krisp should only trigger on events with Zoom/Meet/Teams links — if it's recording events without them, contact support).