# Threadline

> Persistent context layer for AI agents. Two lines of code give any agent persistent, user-owned memory across every model, session, and product.

Threadline is a developer infrastructure product. It exposes an SDK and an MCP server. `tl.inject()` enriches a prompt with relevant user context before an LLM call; `tl.update()` writes new facts after the response. Memory is user-owned with scoped grants, audit trails, and hard delete.

API base URL: https://app.threadline.to — all endpoints are under /api/context/*, authenticated with `Authorization: Bearer tl_live_...`.

## Docs

- [Quickstart](https://threadline.to/docs): Install, authenticate, your first run (update() then inject()), and the steady-state loop. TypeScript and cURL.
- [Response reasons](https://threadline.to/docs/reasons): Every inject() reason value — empty, no_grant, no_scopes, empty_capsule, revoked — plus update()'s denied_scopes and notice fields.
- [Governance](https://threadline.to/docs/governance): The seven scopes, how grants work, write-time enforcement of denials, and scope-filtered reads.
- [SDK reference](https://threadline.to/docs/sdk): TypeScript SDK methods — get, inject, update, grant, revoke.
- [REST API reference](https://threadline.to/docs/api-reference): HTTP endpoints for inject, update, read, delete, and health.
- [MCP setup](https://threadline.to/docs/mcp-setup): One-line install of the Threadline MCP server for Cursor and Claude Desktop.

## Pages

- [Home](https://threadline.to/): Positioning, free-tier signup, two-line code example.
- [How it works](https://threadline.to/how-it-works): Six-step walkthrough from signup to first inject() call.
- [Use cases](https://threadline.to/use-cases): Production patterns from agents shipping with Threadline.
- [Pricing](https://threadline.to/pricing): Free $0/10K memories, Builder $49/100K, Scale $299/1M.
- [Comparison](https://threadline.to/comparison): Threadline vs Mem0 vs Supermemory feature matrix.
- [Threadline vs Mem0](https://threadline.to/vs/mem0): Side-by-side, focused on pricing and governance.
- [Threadline vs Supermemory](https://threadline.to/vs/supermemory): Retrieval and grants comparison.
- [Threadline vs Zep](https://threadline.to/vs/zep): Hosted vs self-hosted tradeoffs.
- [Alternatives to Mem0](https://threadline.to/alternatives-to-mem0): Roundup of memory infrastructure options in 2026.

## Guides

- [Guides index](https://threadline.to/guides): Industry-specific integration recipes.
- [AI coding assistant](https://threadline.to/guides/ai-coding-assistant)
- [Customer support agent](https://threadline.to/guides/customer-support-agent)
- [Voice agent](https://threadline.to/guides/voice-agent)
- [Multi-agent handoff](https://threadline.to/guides/multi-agent-handoff)
- [Sales assistant](https://threadline.to/guides/sales-assistant)
- [Tutor / learning app](https://threadline.to/guides/tutor-learning-app)
- [Writing tool](https://threadline.to/guides/writing-tool)
- [Financial advisor](https://threadline.to/guides/financial-advisor)
- [Health and fitness coach](https://threadline.to/guides/health-fitness-coach)
- [Onboarding flow](https://threadline.to/guides/onboarding-flow)
- [Recruiting assistant](https://threadline.to/guides/recruiting-assistant)

## Blog

- [Blog index](https://threadline.to/blog): Essays on memory infrastructure for AI agents.
- [Why every AI agent suffers from amnesia](https://threadline.to/blog/why-ai-agents-forget)
- [Four things were wrong with my own permission layer](https://threadline.to/blog/four-things-wrong-with-my-permission-layer)

## Optional

- [Changelog](https://threadline.to/changelog): Release notes.
- [Security](https://threadline.to/security): Encryption, grants, audit, infrastructure.
- [Status](https://threadline.to/status): Live system status.
- [Privacy](https://threadline.to/privacy)
- [Terms](https://threadline.to/terms)
- [Cookies](https://threadline.to/cookies)