Every customer conversation,remembered.
Voice bots, chat bots, and support agents forget every caller the moment a session ends. ConvoMem gives them durable customer memory — so the next interaction starts with context, not repetition.
- 4
- channels unified
- 20
- memories / recall
- 3
- official SDKs
- 1
- customer timeline
Run a message through the real pipeline.
Paste a message or drop a small CSV. It runs the same capture, scoring, redaction, reconciliation, and insight pipeline we use in production — streamed back live. Nothing is stored.
extract→filter→score→redact→reconcile→insights
What it feels like for your customer.
Three real moments — a returning caller, a frustrated customer, an expansion. In each one, the agent already knows the story.
Customer intelligence,
assembled automatically.
Available in v1.2.0, including the official TypeScript, Python, and Rust SDKs and the hosted MCP endpoint.
- 01
Memory capture
Context-aware facts are grounded to source turns, attributed to the right speaker, and processed through a pollable async capture job.
voice · chat · sms · email - 02
Identity resolution
One person, one profile. Conversations are linked across channels by phone, email, or your own customer ID.
cross-channel - 03
Controlled recall
Semantic, keyword, temporal, and lifecycle signals rank a bounded context block with caller-set limits and optional diagnostics.
20 memories · 800 tokens - 04
Memory lifecycle
Versions, provenance, expiry, supersession, reviewable patterns, and commitment state stay visible without leaking stale rows into recall.
auditable - 05
PII redaction
Mandatory secrets are dropped, while the configured agent policy can mask or reject recognized sensitive memory candidates before persistence.
policy controlled - 06
Analytics & insights
Ask questions in plain language. Get dashboards, charts, and answers grounded in your own data.
natural language - 07
Integration surfaces
Use the REST API, hosted MCP tools, webhooks, embeds, or the official TypeScript, Python, and Rust SDKs without exposing operator-only APIs.
rest · mcp · sdk - 08
Lead Playbooks
Turn conversations into qualified leads automatically — define a playbook and matching opportunities are detected, scored, and surfaced across every channel.
ai-driven
Three steps. Zero friction.
Capture
A single API call sends a message in. ConvoMem extracts the facts worth keeping and redacts the rest.
Reconcile
New facts are matched to the right customer, deduplicated, and contradictions are resolved automatically.
Inject
On the next turn, the relevant memories are returned as context — so the agent already knows the story.
Transparent. No surprises.
Starter
For small teams
- 10,000 customers
- 500K API calls/mo
- 250 AI chat queries/mo
- 5 team members
- 5 webhooks
- 2 agents
- 365-day retention
Growth
For growing platforms
- 75,000 customers
- 2M API calls/mo
- 1,000 AI chat queries/mo
- 20 team members
- 20 webhooks
- 5 agents
- 5 active lead playbooks
- 3-year retention
Pro
For scaling AI products
- 150,000 customers
- 5M API calls/mo
- 3,000 AI chat queries/mo
- 40 team members
- 40 webhooks
- 15 agents
- 15 active lead playbooks
- Entity knowledge graph
- Custom PII filters
- 5-year retention
Scale
Capacity-planned for enterprise AI products
- 500,000 customers
- 15M API calls/mo
- 10,000 AI chat queries/mo
- 100 team members
- 100 webhooks
- 50 agents
- 50 active lead playbooks
- Entity knowledge graph
- Custom PII filters
- 10-year retention
Custom
Contact sales
- Unlimited everything
- On-prem deployment on request
- Custom SLA
- Dedicated Discord channel
Starter, Growth, and Pro start with a free 7-day trial — no card required. Scale is sales-led so we can size capacity, retention, support, and deployment requirements before activation.
Questions, answered.
Every conversation remembered · from the first call
Stop introducing your AI to the same customer twice.
Start your 7-day free trial. No credit card. On-premises deployment available on request.