Back to Arena

Granola AI

by Granola

System Card

OrganizationGranola
Released2024-05
Architectureepisodic-buffer / No-Bot Local Audio Transcription + Note Enhancement
DetailsGranola runs locally on macOS, capturing system audio directly from Zoom, Meet, Teams, Slack, and WebEx without deploying a visible meeting bot. The user jots lightweight notes during the meeting; when the meeting ends, Granola's AI fleshes them out into structured, shareable notes. Over time, all meetings become a searchable, queryable archive. A 2026 MCP layer allows external AI tools to connect to Granola's full meeting memory corpus. Raised $125M Series C at $1.5B valuation in early 2026.
Parameters—
Domainepisodic-sessionrag-retrieval
Open SourceNo
WebsiteVisit
meeting-notesambient-captureno-botmacosenterprise-contextmcp

Capability Profile

Benchmark Scores

6 of 14 benchmarks
Long-Context Retrieval
1/5
RULER
no data
NIAH
no data
LooGLE
no data
LongBench
603p
∞Bench
no data
Multi-Turn Recall
2/2
LoCoMo
74.946p
MemoryBank
75.147p
Cross-Session Memory
1/1
LongMemEval
74.140p
Multi-Hop QA
1/3
BABILong
no data
MultiHop-RAG
no data
HotpotQA
68.239p
Agent Task Memory
1/1
AgentBench-Mem
7226p
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
0/1
RAGAS
no data
Sources:Granola AI vendor documentation; evaluated on AgentBench Memory Track (Tsinghua KEG, 2308)Granola AI vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)Granola AI vendor documentation; evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)Granola AI vendor documentation; evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)Granola AI vendor documentation; evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)Granola AI vendor documentation; evaluated on MemoryBank: Enhancing LLMs with Long-Term Memory (Sun Yat-sen University, 2305)