> **Agent?** Fastest path: MCP at `https://api.mnemom.ai/mcp` — call `get_started` first (zero-auth, no args). Full agent guide: # What's new | Mnemom # Trust, for the agentic internet. An independent record of what AI agents did, and proof you can trust it. New · Oct 2026 ## Two apps in invite-only preview: Research and Mnemom Agent - researchInvite-only previewBuilds a deep report on a company or person, with each claim linked to its source. [Learn more](#research)[Join the waitlist](/waitlist?interest=research) - mnemom agentInvite-only previewSee exactly what your coding agent did, with a record you can show others. One command to start (Claude Code today). [Learn more](#mnemom-agent)[Join the waitlist](/waitlist?interest=mnemom-agent) Research ## Deep research reports, with sources you can check. Research writes a deep report on a person or company for a named reader. Statements are traced to their sources so a reader can check them. - Claims link to sourcesThe trail is kept so a reader can follow it. - Ready to exportExport a finished report as a PDF or Word file. - Saves research timeResults depend on the topic and sources. Mnemom Agent ## Know what your coding agent did, with a record to show for it. Run mnemom agent instead of starting your coding agent directly. Every session is recorded under its own identity, so you can review what it did and share the record with others. - See every sessionReview what your agent did in your traces. - A record you can shareShow others what ran, not just your word. - Nothing to rewriteOne command, no change to your agent. Claude Code today; Anthropic only. ![](/images/mnemom-agent-video-poster.webp) **The trust plane** checks, signs and records every agent action on any model - Checks - Signs - Records [Explore](/platform) ## Put your agent on the record. [Talk to sales](/contact) --- _Source: /whats-new/index.html · Generated by build-markdown-mirrors.mjs · For agent-readability commitment #4 see https://www.mnemom.ai/for-agents/_