Snipara is the Project Brain for AI coding agents.
Its shared Project Intelligence layer gives Copilot, Claude Code, Cursor, Codex, and other MCP clients the project's reviewed decisions, active work, code impact, verification evidence, and handoffs before they edit. Your agent keeps using its own LLM; Snipara keeps project context, memory, and workflow continuity attached to the project across sessions, users, tools, and model switches.
Snipara turns an editor install into a live project context in under a minute. Cursor uses the create-snipara bootstrap; VS Code can run the same activation as a native workspace command.
VS Code
|
v
Snipara: Activate Workspace
|
v
Editor Activation Contract
|
v
Project Brain
|
+-- Work Brief
+-- Impact
+-- Verification
+-- Receipts
Prepared extension release: 2.0.9.
Available from the VS Code Marketplace and Open VSX Registry for VS Code, Cursor, VSCodium, and other VS Code-compatible editors.
- Runs the native VS Code entry point for Snipara's shared activation engine: install extension, activate workspace, consume the editor activation manifest, open a First Work Brief, then hand off to Copilot.
- Activates the open workspace by launching
create-snipara@latest init --client vscode --starter --json, reading.snipara/activation/activation-manifest.json, and rendering the versioned editor contract. - Runs a no-account demo against Snipara's project context as a fallback. The demo is opt-in and limited to 3 queries.
- Connects VS Code-compatible editors to a Snipara project through API credentials stored in VS Code SecretStorage.
- Lets users query source-backed project context, upload documentation, inspect memories, review decisions, and monitor index health.
- Exposes 64 Command Palette actions across project context, memory, documents, orchestration, sandbox execution, swarm coordination, and local readiness.
- Exposes 38 GitHub Copilot Language Model Tools in VS Code agent mode.
- Registers a Snipara MCP Server Definition Provider for VS Code's MCP integration when the host supports it.
- Keeps Cursor users supported through the extension UI plus the separate hosted MCP setup for Cursor chat.
- Stop re-explaining the project to every coding agent.
- Reuse reviewed decisions and source-backed context.
- Know what files and risks matter before editing.
- Start every session with a First Work Brief.
Before Snipara:
Implement OAuth.
The agent starts searching the workspace from scratch.
After Snipara:
Snipara: Activate Workspace
Opening First Work Brief...
✓ Decisions
✓ Active work
✓ Architecture
✓ Impact
✓ Verification
Snipara Active.
VS Code keeps the native editor workflow. Snipara supplies the activation manifest, First Work Brief, Copilot handoff, and source-backed tools the agent can use before changing files.
Snipara fits the VS Code workflow without sending users to a separate dashboard:
- session starts with Snipara: Activate Workspace
- activation output is rendered from
.snipara/activation/activation-manifest.json - Copilot handoff copies the Work Brief context into chat
- status bar shows
Snipara Activeafter activation - commands and Copilot Language Model Tools keep retrieval, memory, verification, and handoff available during work
The primary first run is your own workspace.
- Install Snipara from VS Code Marketplace or Open VSX.
- Open a project that has docs such as
README,docs/**,AGENTS.md,CLAUDE.md, or changelogs. - Click Activate Workspace / Build my First Work Brief.
- Sign in with GitHub if prompted; Snipara stores credentials in VS Code SecretStorage.
- Snipara runs the shared activation engine, opens a source-backed First Work Brief from the activation manifest, and offers a Copilot handoff prompt.
This is the VS Code-native version of the create-snipara activation flow. The activation engine stays shared across Snipara surfaces; VS Code adds editor-native UX with commands, SecretStorage, panels, status bar, manifest rendering, and Copilot handoff.
Snipara keeps the activation logic centralized instead of duplicating it in every integration.
| Surface | Entry Point | What Users Get |
|---|---|---|
| VS Code extension | Native Snipara: Activate Workspace command | Sign-in when needed, create-snipara --json, activation manifest rendering, First Work Brief panel, Copilot handoff |
| Cursor plugin | create-snipara bootstrap |
Project-local Cursor rules, activation artifacts, First Work Brief path, hosted MCP upgrade with SNIPARA_API_KEY |
create-snipara CLI |
npx create-snipara@latest init --client <client> --starter |
Shared activation package for Cursor, Claude Code, Codex, VS Code-compatible clients, and generic MCP clients |
The product promise is the same across surfaces: install Snipara, open your project, and give the agent live project context before it starts work.
| Editor | Install Source | Supported Surfaces |
|---|---|---|
| VS Code | VS Code Marketplace | Commands, sidebar views, status bar, file explorer upload, Copilot Language Model Tools, MCP Server Definition Provider |
| Cursor | Open VSX | Commands, sidebar views, status bar, file explorer upload, keyboard shortcut. Configure Snipara MCP separately for Cursor chat |
| VSCodium / Theia / Open VSX editors | Open VSX | Commands, sidebar views, file explorer upload, sandbox/local readiness where host APIs allow |
Cursor and other Open VSX editors do not expose VS Code's GitHub Copilot-specific Language Model Tool APIs. For AI chat access in Cursor, configure the hosted MCP integration from Snipara Cursor docs.
- Workspace activation: The first-run path runs the shared
create-sniparaeditor activation contract, opens a source-backed First Work Brief, and offers direct Copilot handoff actions. - Local project proof: The First Work Brief is backed by
.snipara/activation/activation-manifest.json, including activation lanes, artifacts, next actions, and source-backed starting points. - Opt-in demo mode: Run 3 demo queries against Snipara's own indexed project context without signing in.
- Instant first demo: The default demo query uses embedded fallback data, so it works even when the network is unavailable.
- Live follow-ups: Custom demo questions call the read-only demo project and fall back silently if the API is unreachable.
- Guided walkthrough: First-run users see the Getting Started flow, with workspace activation first and the demo kept as fallback.
- Manifest-backed UX: VS Code consumes
surfaceContract,artifacts.firstBriefPath,artifacts.handoffPath,lanes, andnextActionsinstead of reimplementing repository scanning.
- Natural-language project context queries with keyword, semantic, or hybrid search.
- Source-backed retrieval with cited files, sections, relevance scores, and token budgets.
- Raw chunk, raw document, line-range, and project-load commands for deeper inspection.
- Multi-query, multi-project query, query decomposition, planning, and orchestration workflows.
- Code graph lookup for neighbors, callers, imports, importers, and structural paths.
- Store and recall facts, decisions, learnings, preferences, todos, and context.
- Use Remember If Novel to avoid duplicate durable memories.
- Run End-of-Task Commit to persist reusable outcomes at the end of a task.
- Browse memories in the sidebar and inspect memory health, review queues, stale entries, and candidate memories.
- Query structured project decisions directly.
- Index health, recommendations, reindexing, search analytics, and query trends.
- Shared context collections and prompt templates for team standards.
- Summary storage for generated document summaries.
- Swarm coordination commands for multi-agent work: swarms, tasks, resource claims, shared state, and broadcast events.
- Native Doctor / Local Readiness view for workspace
.envfiles, provider keys, Snipara auth, Snipara Sandbox, and Docker. - Sandbox commands for Docker-isolated execution, local execution, logs, and trajectory visualization.
- Local setup guidance installs the PyPI package
snipara-sandbox. - Core extension features do not require
snipara-companion.
- Snipara activity bar with Welcome, Results, Session Context, Memories, Doctor / Local Readiness, and Swarm Dashboard views.
- Status bar entry for sign-in or quick query.
Cmd+Shift+R/Ctrl+Shift+Rshortcut for Snipara: Ask Question.- File explorer upload for
.md,.mdx, and.txtfiles. - Workspace-first document sync: single-folder workspaces sync without an unnecessary folder picker; multi-root workspaces use a VS Code quick pick.
- Optional auto-restore and auto-save of session context.
All 64 commands are available from the Command Palette under Snipara, Snipara Sandbox, and Snipara Doctor.
| Command | Purpose |
|---|---|
| Snipara: Activate Workspace | Run create-snipara@latest init --client vscode --starter --json, read the activation manifest, and open a First Work Brief |
| Snipara: Ask Question | Query source-backed project context. Falls back to demo mode when not signed in |
| Snipara: Search Project Context | Search indexed project context by pattern. Falls back to demo mode when not signed in |
| Snipara: Try Demo Query | Run the opt-in demo query |
| Snipara: Multi-Query | Run multiple retrieval queries in parallel |
| Snipara: Multi-Project Query | Search across accessible team projects |
| Snipara: Decompose Query | Break a complex question into sub-queries |
| Snipara: Generate Plan | Build an execution plan for context exploration |
| Snipara: Show Current Context | Load active session context |
| Snipara: Inject Context | Add selected text or custom text to session context |
| Snipara: Clear Context | Clear active session context |
| Command | Purpose |
|---|---|
| Snipara: Remember | Store a fact, decision, learning, preference, todo, or context memory |
| Snipara: Recall Memory | Search durable project memories |
| Snipara: Browse Memories | Browse memories in the sidebar |
| Snipara: Refresh Memories | Refresh the Memories view |
| Snipara: Forget Memory | Delete a memory after confirmation |
| Snipara: End-of-Task Commit | Persist reusable task outcomes into memory |
| Snipara: Show Memory Health | Inspect duplicates, stale memories, and review signals |
| Snipara: Show Memory Review Queue | Inspect candidate, rejected, stale, or active review items |
| Snipara: Query Decisions | Search structured project decisions |
| Command | Purpose |
|---|---|
| Snipara: Upload Document | Upload a .md, .mdx, or .txt document |
| Snipara: Sync Documents | Bulk sync workspace documentation |
| Snipara: Upload Shared Document | Upload a document into a shared context collection |
| Snipara: Load Shared Context | Load shared team context |
| Snipara: List Collections | Browse shared context collections |
| Snipara: Browse Templates | List reusable prompt templates |
| Snipara: Apply Template | Load a prompt template by slug |
| Snipara: Store Summary | Store an AI-generated document summary |
| Snipara: View Summaries | Browse stored summaries |
| Snipara: Delete Summary | Delete a stored summary |
| Snipara: Show Statistics | View indexed files, sections, lines, and token counts |
| Snipara: Show Index Health | Inspect coverage, quality, tiers, and stale content |
| Snipara: Show Index Recommendations | Get index improvement recommendations |
| Snipara: Reindex Project | Trigger document or code index maintenance |
| Snipara: Show Search Analytics | View query performance and tool usage analytics |
| Command | Purpose |
|---|---|
| Snipara: Show Section | Display a retrieved source section |
| Snipara: Get Chunk by ID | Load cited source content by chunk ID |
| Snipara: Read Lines | Read a specific line range from indexed documentation |
| Snipara: Code Graph Lookup | Inspect code graph neighbors, callers, or imports |
| Snipara: Load Document (Raw) | Load raw document content by path |
| Snipara: Load Project | Load a broader project context dump with path filtering |
| Snipara: Orchestrate | Run multi-round scan, search, and raw-load exploration |
| Snipara: Build REPL Context | Package project context with Python helpers for sandbox sessions |
| Command | Purpose |
|---|---|
| Snipara: Create Swarm | Create an agent swarm |
| Snipara: Join Swarm | Join an existing swarm |
| Snipara: Claim Resource | Lock a file, function, module, or resource |
| Snipara: Release Resource | Release a resource claim |
| Snipara: Create Task | Add a task to a swarm queue |
| Snipara: Claim Task | Claim the highest-priority available task |
| Snipara: Complete Task | Mark a task as completed or failed |
| Snipara: Broadcast Event | Send an event to swarm agents |
| Snipara: Get Shared State | Read shared swarm state by key |
| Snipara: Set Shared State | Write shared swarm state with optimistic locking |
| Command | Purpose |
|---|---|
| Snipara: Configure | Sign in with GitHub or configure API credentials manually |
| Snipara: Show Project Settings | View current Snipara project settings |
| Snipara Sandbox: Execute in Docker (Isolated) | Run code through Snipara Sandbox with Docker isolation |
| Snipara Sandbox: Execute Locally | Run code through Snipara Sandbox locally |
| Snipara Sandbox: View Execution Logs | Browse sandbox execution logs |
| Snipara Sandbox: Launch Trajectory Visualizer | Open the trajectory visualizer |
| Snipara Doctor: Refresh Local Readiness | Re-run local setup checks |
| Snipara Doctor: Open Env File | Open or create a workspace .env file |
| Snipara Doctor: Use Workflow | Pick a context workflow based on current readiness |
| Snipara Doctor: Use Sandbox | Pick a sandbox action based on detected Snipara Sandbox/Docker status |
| Snipara Doctor: Run Sandbox Doctor | Run snipara-sandbox doctor when available |
In VS Code agent mode, Snipara contributes 38 Language Model Tools covering:
- project context retrieval, search, quick ask, multi-query, planning, decomposition, and multi-project search;
- memory storage, recall, browsing, deletion, novelty checks, memory health, memory review, and end-of-task commits;
- shared context, document upload, raw document load, full project load, orchestration, and REPL context packaging;
- source chunk loading, code graph neighbors, callers, imports, and shortest paths;
- structured decisions, index health, recommendations, reindexing, search analytics, and query trends;
- hierarchical task inspection, recommendations, metrics, and Python execution through Snipara Sandbox.
The extension also declares a Snipara MCP Server Definition Provider for VS Code hosts that support MCP server definitions. Hosted MCP availability depends on the editor, VS Code version, project configuration, and Snipara server deployment.
| Setting | Default | Description |
|---|---|---|
snipara.apiKey |
empty | Legacy/manual API key setting. Sign-in stores keys in VS Code SecretStorage |
snipara.projectId |
empty | Snipara project ID or slug |
snipara.serverUrl |
https://api.snipara.com |
Snipara API server URL |
snipara.maxTokens |
4000 |
Max context budget for retrieval queries |
snipara.searchMode |
hybrid |
Search mode: keyword, semantic, or hybrid |
snipara.enableAutoRestore |
false |
Restore previous session context on startup |
snipara.enableAutoSave |
false |
Save session context periodically and on shutdown |
snipara.sandboxEnabled |
true |
Enable Snipara Sandbox status and commands |
- VS Code 1.93.0+, Cursor 0.40+, VSCodium, or another compatible Open VSX editor.
- No account is required for the opt-in demo.
- A free Snipara account is required to query and manage your own project context.
- Optional for sandbox execution: snipara-sandbox and Docker for isolated runs.
- Optional for companion workflows:
snipara-companion.
pnpm install
pnpm run type-check
pnpm run compile
pnpm run packagePackaging uses vsce package --no-dependencies. Publishing targets:
pnpm run publish:vsce
pnpm run publish:ovsxGenerated .vsix files are local release artifacts and are ignored by git.