Fireflies cost tuning
'Optimize Fireflies.ai subscription costs through seat auditing, selective recording, and plan sizing.From its SKILL.md
npx -y skills add jeremylongshore/claude-code-plugins-plus-skills --skill fireflies-cost-tuningAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
What its file declares
Copied from the file, not written here
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
7.0 KB, ~1.6k tokens by cl100k_base, as published. Nobody here has run it
Fireflies.ai Cost Tuning
Overview
Optimize Fireflies.ai subscription costs. Fireflies charges per-seat per month. The main levers: remove unused seats, configure selective recording, manage storage, and right-size your plan tier.
Pricing Reference
| Plan | Price | API Access | Key Features |
|---|---|---|---|
| Free | $0 | 50 req/day | 800 min storage, limited transcription |
| Pro | ~$18/seat/month | 50 req/day | 8,000 min/seat, AI summaries |
| Business | ~$29/seat/month | 60 req/min | Unlimited transcription, CRM, analytics |
| Enterprise | Custom | 60 req/min | SSO, Super Admin webhooks, custom |
Instructions
Step 1: Audit Seat Utilization via API
set -euo pipefail
# List all workspace users with their transcript counts
curl -s -X POST https://api.fireflies.ai/graphql \
-H "Authorization: Bearer $FIREFLIES_API_KEY" \
-H "Content-Type: application/json" \
-d '{"query": "{ users { name email user_id num_transcripts minutes_consumed } }"}' \
| jq '.data.users | sort_by(.num_transcripts) | .[] | {name, email, transcripts: .num_transcripts, minutes: .minutes_consumed}'
// Automated utilization report
async function seatUtilizationReport() {
const data = await firefliesQuery(`{
users {
name email user_id
num_transcripts minutes_consumed
recent_meeting
}
}`);
const users = data.users;
const inactive = users.filter((u: any) => u.num_transcripts < 2);
const active = users.filter((u: any) => u.num_transcripts >= 2);
console.log(`Total seats: ${users.length}`);
console.log(`Active (2+ transcripts): ${active.length}`);
console.log(`Inactive (<2 transcripts): ${inactive.length}`);
console.log(`Potential savings: ${inactive.length} seats * $29/mo = $${inactive.length * 29}/mo`);
if (inactive.length > 0) {
console.log("\nInactive seats to review:");
for (const u of inactive) {
console.log(` ${u.email}: ${u.num_transcripts} transcripts, last meeting: ${u.recent_meeting || "never"}`);
}
}
return { total: users.length, active: active.length, inactive, savings: inactive.length * 29 };
}
Step 2: Configure Selective Recording
Instead of recording every meeting, configure auto-join rules in Fireflies Settings > Auto-Join:
# Recommended recording policy
record_always:
- External meetings (client/prospect calls)
- Meetings with 3+ participants
- Meetings with keywords: "review", "planning", "standup", "demo"
skip_recording:
- 1-on-1 informal chats
- Social events
- Meetings shorter than 5 minutes
- Recurring "lunch" or "coffee" meetings
Estimated savings: Teams recording every meeting typically waste 30-50% of transcription credits on low-value meetings.
Step 3: Manage Storage to Avoid Forced Upgrades
// Check storage and clean up old transcripts
async function storageAudit() {
const data = await firefliesQuery(`{
transcripts(limit: 100) {
id title date duration
}
user { minutes_consumed }
}`);
const now = Date.now();
const transcripts = data.transcripts;
// Find transcripts older than 90 days
const old = transcripts.filter((t: any) => {
const age = (now - new Date(t.date).getTime()) / 86400000;
return age > 90;
});
console.log(`Total transcripts: ${transcripts.length}`);
console.log(`Older than 90 days: ${old.length}`);
console.log(`Minutes consumed: ${data.user.minutes_consumed}`);
return { total: transcripts.length, old, minutesUsed: data.user.minutes_consumed };
}
// Delete old transcripts to free storage
async function deleteOldTranscripts(ids: string[]) {
for (const id of ids) {
await firefliesQuery(`
mutation($id: String!) {
deleteTranscript(transcript_id: $id)
}
`, { id });
console.log(`Deleted: ${id}`);
// deleteTranscript is rate limited: 10/min
await new Promise(r => setTimeout(r, 6500));
}
}
Step 4: Right-Size Your Plan
# Decision matrix
choose_pro:
when:
- Team averages <15 meetings/week per person
- No CRM integration needed
- Basic AI summaries sufficient
saves: $11/seat/month vs Business
choose_business:
when:
- Sales team recording every call
- CRM integration required (Salesforce, HubSpot)
- Meeting analytics dashboards needed
- 20+ meetings/week per person
choose_enterprise:
when:
- SSO/SAML required
- Super Admin webhooks needed (org-wide meeting data)
- Custom data retention policies
- Dedicated support
Step 5: API Cost Optimization
// Free/Pro plans: 50 requests/day. Make every request count.
// Strategy: Fetch meeting list once, cache aggressively
async function efficientDailySync() {
// One request: get all recent transcripts
const data = await firefliesQuery(`{
transcripts(limit: 50) {
id title date duration
summary { overview action_items }
}
}`);
// Process locally -- no additional API calls
const today = new Date().toDateString();
const todaysMeetings = data.transcripts.filter(
(t: any) => new Date(t.date).toDateString() === today
);
console.log(`Today's meetings: ${todaysMeetings.length}`);
console.log(`API requests used: 1 of 50 daily budget`);
return todaysMeetings;
}
Cost Savings Summary
| Lever | Typical Savings |
|---|---|
| Remove inactive seats | $29/seat/month |
| Selective recording | 30-50% fewer transcriptions |
| Pro vs Business downgrade | $11/seat/month |
| Storage cleanup | Avoid forced tier upgrade |
| API request caching | Stay within Free/Pro limits |
Error Handling
| Issue | Cause | Solution |
|---|---|---|
| Can't remove seat | User has admin role | Reassign admin first |
| Storage limit warning | Too many transcripts | Delete old transcripts |
| API daily limit hit | Free/Pro 50 req/day | Cache results, batch efficiently |
| Unexpected invoice increase | Auto-provisioned members | Disable auto-provisioning |
Output
- Seat utilization report with inactive members identified
- Selective recording policy configured
- Storage audit with cleanup recommendations
- Plan right-sizing recommendation
Resources
Next Steps
For architecture design, see fireflies-reference-architecture.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.