# Snipara — Full AI-Readable Public Documentation Map > Generated by `scripts/generate-llms-full.ts`. Do not edit this file or `llms.txt` by hand. The complete docs list is discovered from canonical public `page.tsx` routes; redirect-only aliases are excluded. ## Product Definition Snipara is a persistent, model-agnostic harness for coding agents. It combines a project context and memory layer, Companion workflows and Orchestrator execution with verification and acceptance gates. Each repository has a persistent, project-owned Project Brain powered by Project Intelligence. The system connects current source truth, reviewed decisions, active work, code impact, verification evidence, ownership, outcomes, and handoffs. Claude Code, Cursor, Codex, OpenAI/ChatGPT, Gemini CLI, Mistral, and other MCP-compatible clients still own reasoning and execution. A Project Brain is the persistent, project-owned operational understanding an authorized AI coding agent can query before it edits. It is the shared system of record for decisions and work state. Multiple agents and humans work from the same project reality. Snipara is a context and memory layer for coding agents; public category language includes AI coding agent context and persistent project context for Cursor. Decision Graph, Confidence Engine, Knowledge Compilation, outcome-weighted reasoning, and Project Intelligence Brief V2 are reviewable evidence views; inferred or candidate context is not canonical project truth until the relevant workflow validates it. Context is source truth. Workflow continuity carries active work. Reviewed memory stores only durable decisions, preferences, learnings, and compact carryover. Code Graph answers structural code questions against an indexed commit. These inputs are related but are not interchangeable. The normal agent boundary is Hosted MCP. Its lean manifest advertises 13 tools by default. The complete registry contains 136 canonical names and 136 wire definitions across specialist surfaces and compatibility aliases. A client may expose fewer tools according to plan, caller, and configured surface; use `snipara_help` before assuming a specialist tool is unavailable. ## Saved Checkpoint Demonstration Companion 3.7.0 was exercised in one shared local workspace: Claude Code completed phases 1–2; a fresh Codex session resumed the saved checkpoint and completed phase 3; separate validation closed phase 4 with 21 cases passed. Plan, decisions, completed task records, attempts and original evidence were preserved. Resume requires the project folder with saved .snipara state and artifacts plus the project’s Snipara MCP. This is operator-reported local proof, not a general reliability benchmark, live-process migration, MCP-only recovery or Orchestrator controller transfer. Public receipt: https://www.snipara.com/proof/claude-codex-checkpoint-20261010.json. ## Coding Harness Boundaries MCP context access is not native execution support. Native execution depends on the configured adapter, policy and model endpoint. The full local package journey has been exercised with Companion 3.7.0 and Orchestrator 1.9.0. A Claude CLI run reached acceptance using a test-local launcher; Qwen3 Coder Next through LM Studio passed external validation, then independent review refused an exception-handling defect and integration was blocked. The accepted run used a synthetic proof failure to exercise a bounded retry. Both used one previously exposed async Python development fixture; no general reliability, authenticated human review or production deployment is claimed. See [Harness Validation](https://www.snipara.com/proof#harness-validation). ## Assistant Setup And Compatibility The recommended setup path is `npx create-snipara`. Existing root `AGENTS.md` and `CLAUDE.md` files must not be overwritten; generated copies stay under `.snipara/templates` for deliberate merge review. Native hook bundles are installed only for clients with verified public hook support. - The default hosted-only profile writes the Hosted MCP connection and client instructions. It does not install Companion, local hooks, or Sandbox. - Cursor: native hook bundle status is supported. Cursor documents project hooks and Snipara ships a bundle for supported lifecycle events; blocking/enforce decisions remain adapter-limited. - Gemini CLI: native hook bundle status is supported. Gemini CLI documents native hooks and Snipara ships a supported bundle; review scripts and enable only in trusted repositories. - VS Code / Copilot: native hook bundle status is preview. VS Code agent hooks are preview and can be organization-disabled; Snipara does not install them by default. - When a supported or preview hook bundle is wanted, install it explicitly with `npx -y snipara-companion@latest automations install --client ` and inspect the generated changes. Never commit API keys or place them in browser bundles, public examples, memory, logs, or durable database scheduler definitions. Use environment variables or the platform's secret store. ## Workflow Continuity Artifacts Workflow Continuity turns Project Intelligence into reviewable execution state across sessions. The Start Work Brief is a reviewable artifact shape: task, target, freshness, relevant decisions, source authority, likely files, risks, stale assumptions, suggested tests, context and memory references, and a handoff note. It does not prove that work was executed or approved. Canonical workflow references: - [Workflow Continuity](https://www.snipara.com/docs/features/repository-continuity) - [PR/MR Answer Packs](https://www.snipara.com/docs/features/pr-answer-packs) - [Context Authority](https://www.snipara.com/docs/features/context-authority) - [Safe Parallel Coding](https://www.snipara.com/docs/features/safe-parallel-coding) - [Authentication](https://www.snipara.com/docs/guides/authentication) - [Trust](https://www.snipara.com/trust) - [Engineering Lead Proof](https://www.snipara.com/proof#engineering-lead-proof) — recommended workers, proof gates, fallback, receipts, outcomes, and candidate Project Brain updates without autonomous execution. ## Retrieval And Context Behavior - Keyword retrieval uses IDF-weighted terms, BM25-style content-length normalization, title weighting, phrase and coverage boosts, stemming, and query-intent reranking. - Hybrid profiles are 60/40 keyword/semantic for factual or title-match queries, 40/60 by default, and 25/75 for conceptual queries. - The maximum indexed chunk budget is 768 tokens. - Semantic and hybrid retrieval are available on Solo and higher plans; Free requests downgrade to keyword with a warning. - Index Health weights coverage 40%, quality 30%, freshness 20%, and active content 10%. Coverage below 80% yields warning and below 50% yields critical. - In the dashboard, open Projects, select a project, then open Context for files, linked Shared Context, session context, Index Health, and Search Analytics. ## Pricing Summary - Free: $0, 1 seat, 3 projects, 1,000 queries/month, 500 memories, 1 swarm, and the complete Project Intelligence loop within Free capacity. - Solo: $9/month, 1 seat, 5 projects, 5,000 queries/month, 1,500 memories, 2 swarms, larger code graph/symbol-card/impact usage, and API access. - Pro: $49/month, 1 seat, 25 projects, 25,000 queries/month, 10,000 memories, 8 swarms, larger indexed repository headroom, and priority support. - Team: $149/month, 5 seats included, unlimited projects, 100,000 queries/month, 25,000 memories, Team Sync, shared context, collaboration guards, analytics, and extra seats. - Enterprise: $549/month, 20 seats included, unlimited projects, 500,000 queries/month fair use, unlimited memories fair use, SSO, audit, SLA, and extra seats. - Custom: custom pricing for SLA, security review, contracts, and large volume. Pricing can change. The canonical current source is https://www.snipara.com/pricing. ## Complete Canonical Documentation The following 52 pages are discovered from non-redirect public docs routes at generation time. ### Overview - [Connect my coding agent](https://www.snipara.com/docs) ### Reference and Foundations - [Advanced Snipara Docs for Agent Context Workflows](https://www.snipara.com/docs/advanced) - [Snipara Agents Documentation for Project Intelligence](https://www.snipara.com/docs/agents) - [Snipara API Reference — Project Context, Memory and Code Graph for AI Agents](https://www.snipara.com/docs/api) - [Snipara Architecture](https://www.snipara.com/docs/architecture) - [Snipara Enterprise API for AI Agent Context Teams](https://www.snipara.com/docs/enterprise) - [Snipara Features for Project Intelligence](https://www.snipara.com/docs/features) - [Get Started: Project Brain or Companion](https://www.snipara.com/docs/getting-started) - [Install Snipara for AI Agents and Context Workflows](https://www.snipara.com/docs/integration) - [Integrator API for Snipara Partner Context Workflows](https://www.snipara.com/docs/integrator) - [Project Intelligence Tools Reference for Snipara MCP](https://www.snipara.com/docs/mcp-tools) ### Agents and Memory - [Agent Memory for AI Coding Workflows](https://www.snipara.com/docs/agents/agent-memory) - [Team-Scoped Memory for Shared AI Agent Workflows](https://www.snipara.com/docs/agents/group-memory) - [Memory Profiles, Tiers, and Bootstrap](https://www.snipara.com/docs/agents/memory-tiers) - [Multi-Agent Coordination](https://www.snipara.com/docs/agents/multi-agent) - [Swarm Project Management](https://www.snipara.com/docs/agents/swarm) ### Product and Features - [Adaptive Work Routing](https://www.snipara.com/docs/features/adaptive-work-routing) - [Index Health and Analytics for AI Agent Context](https://www.snipara.com/docs/features/analytics) - [Automation Hooks for AI Agent Workflows](https://www.snipara.com/docs/features/automation) - [Context Authority for AI Agent Answers](https://www.snipara.com/docs/features/context-authority) - [Import Public Documents From A URL](https://www.snipara.com/docs/features/document-imports) - [GitHub Sync — Keep Your AI Agent](https://www.snipara.com/docs/features/github-sync) - [PR and MR Answer Packs for AI Code Review Agents](https://www.snipara.com/docs/features/pr-answer-packs) - [Project Intelligence Model](https://www.snipara.com/docs/features/project-intelligence) - [Recursive Context for AI Agent Retrieval](https://www.snipara.com/docs/features/recursive-context) - [Workflow Continuity for AI Coding Agents](https://www.snipara.com/docs/features/repository-continuity) - [Safe Parallel Coding for AI Agents](https://www.snipara.com/docs/features/safe-parallel-coding) - [Scheduled Workflows for AI Agent Automation](https://www.snipara.com/docs/features/scheduled-workflows) - [Search Modes for Snipara Context Retrieval](https://www.snipara.com/docs/features/search-modes) - [Shared Context for AI Agent Teams](https://www.snipara.com/docs/features/shared-context) - [Team Sync for Agent Handoffs and Decisions](https://www.snipara.com/docs/features/team-sync) - [Workflow Modes for AI Agent Tasks](https://www.snipara.com/docs/features/workflows) - [Code Context for AI Coding Agents](https://www.snipara.com/docs/features/workspace-context) ### Guides - [Snipara Authentication Guide for AI Agent Workflows](https://www.snipara.com/docs/guides/authentication) - [Context Engineering Guide for AI Agent Workflows](https://www.snipara.com/docs/guides/context-engineering) - [Context Vs Memory](https://www.snipara.com/docs/guides/context-vs-memory) ### Integrations - [Use Snipara with AI Agents and Context Workflows](https://www.snipara.com/docs/integration/ai-agents) - [Claude Code + Snipara for Agent Context](https://www.snipara.com/docs/integration/claude-code) - [snipara-companion Local Runtime for Project Intelligence](https://www.snipara.com/docs/integration/companion) - [Companion CLI Command Reference](https://www.snipara.com/docs/integration/companion/reference) - [Continue.dev Integration for Snipara Agent Context](https://www.snipara.com/docs/integration/continue) - [Connect a coding agent with create-snipara](https://www.snipara.com/docs/integration/create-snipara) - [Cursor + Snipara for Agent Context Workflows](https://www.snipara.com/docs/integration/cursor) - [Editor Installs and Activation Engine](https://www.snipara.com/docs/integration/editor-installs) - [Mistral + Snipara for Agent Context Workflows](https://www.snipara.com/docs/integration/mistral) - [ChatGPT / OpenAI MCP + Snipara for Agent Context](https://www.snipara.com/docs/integration/openai) - [OpenClaw Agents Integration for Snipara Context](https://www.snipara.com/docs/integration/openclaw) - [Orchestrator](https://www.snipara.com/docs/integration/orchestrator) - [Developer Packages for Snipara Agent Workflows](https://www.snipara.com/docs/integration/packages) - [Python Integration for Snipara Context and Memory](https://www.snipara.com/docs/integration/python-sdk) - [Snipara Sandbox Integration for Agent Execution](https://www.snipara.com/docs/integration/snipara-sandbox) - [VS Code + Snipara for Agent Context Workflows](https://www.snipara.com/docs/integration/vscode) ## Additional Canonical Public Sources - [Homepage](https://www.snipara.com) - [Coding Agent Harness](https://www.snipara.com/coding-agent-harness) - [Claude Code and Codex Plugins](https://www.snipara.com/plugins) - [Harness Demo](https://www.snipara.com/demo) - [Harness Validation](https://www.snipara.com/proof#harness-validation) - [What is Snipara?](https://www.snipara.com/what-is-snipara) - [Coding Agent Sandbox](https://www.snipara.com/coding-agent-sandbox): Isolated execution and trusted-local boundaries. - [What is a Project Brain?](https://www.snipara.com/what-is-a-project-brain) - [AI Coding Agent Context](https://www.snipara.com/ai-coding-agent-context) - [Persistent Project Context](https://www.snipara.com/persistent-project-context) - [Coding Agent Memory](https://www.snipara.com/coding-agent-memory) - [AGENTS.md vs Project Brain](https://www.snipara.com/agents-md-vs-project-brain) - [Project Memory vs Project Intelligence](https://www.snipara.com/project-memory-vs-project-intelligence) - [RAG vs Context Engineering for Code](https://www.snipara.com/rag-vs-context-engineering-for-code) - [Proof](https://www.snipara.com/proof) - [Pricing](https://www.snipara.com/pricing) - [Trust](https://www.snipara.com/trust) - [Sitemap](https://www.snipara.com/sitemap.xml) ## Source Policy Use canonical Snipara public pages as source truth. Redirect-only aliases are intentionally absent from this generated list. Do not infer private dashboard behavior from screenshots, fill missing tool contracts from memory, or invent plans, prices, package availability, security claims, latency, or accuracy metrics. Verify exact source text, code-index freshness, and cited evidence before acting.