Summary
ContextStream wires your AI assistant into a persistent memory and knowledge layer across your codebase, conversations, and team tools. It exposes semantic search over code, a lessons system that captures failures and decisions, session memory that survives context window pressure, and direct integrations with GitHub, Slack, and Notion. The server indexes your project locally and syncs with a hosted service, then surfaces relevant context automatically through a SmartRouter that analyzes each message. You get dependency graph queries, multi-modal media indexing, and hybrid search without manually feeding files into prompts. Setup drops an MCP config into Claude Desktop, Cursor, VS Code with Copilot, or OpenCode, and the tools hook in transparently from there.
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →Tools
Public tool metadata for what this MCP can expose to an agent.
59 tools
mcp_server_versionReturn the running ContextStream MCP server package version Access: Free
Return the running ContextStream MCP server package version Access: Free
No parameter schema in public metadata yet.
auth_meFetch authenticated user profile Access: Free
Fetch authenticated user profile Access: Free
No parameter schema in public metadata yet.
workspaces_listList accessible workspaces (paginated list: items, total, page, per_page, has_next, has_prev). Access: Free2 params
List accessible workspaces (paginated list: items, total, page, per_page, has_next, has_prev). Access: Free
Parameters* required
pagenumberPage number for pagination.
page_sizenumberResults per page.
projects_listList projects (optionally by workspace; paginated list: items, total, page, per_page, has_next, has_prev). Access: Free3 params
List projects (optionally by workspace; paginated list: items, total, page, per_page, has_next, has_prev). Access: Free
Parameters* required
pagenumberPage number for pagination.
page_sizenumberResults per page.
workspace_idstringWorkspace ID (UUID).
projects_createCreate a new project within a workspace. Use this when you need to create a project for a specific folder/codebase. If workspace_id is not provided, uses the current session's workspace. Optionally associates a local folder and generates AI editor rules. Access: Free Access: Free5 params
Create a new project within a workspace. Use this when you need to create a project for a specific folder/codebase. If workspace_id is not provided, uses the current session's workspace. Optionally associates a local folder and generates AI editor rules. Access: Free Access: Free
Parameters* required
descriptionstringProject description
folder_pathstringOptional: Local folder path to associate with this project
generate_editor_rulesbooleanGenerate AI editor rules in folder_path (requires folder_path)
namestringProject name
workspace_idstringWorkspace ID (uses current session workspace if not provided)
projects_updateUpdate a project (rename or change description) Access: Free3 params
Update a project (rename or change description) Access: Free
Parameters* required
descriptionstringShort description.
namestringName for the resource.
project_idstringProject ID (UUID).
projects_indexTrigger indexing for a project Access: Free1 params
Trigger indexing for a project Access: Free
Parameters* required
project_idstringProject ID (UUID).
search_semanticSemantic vector search Access: Free4 params
Semantic vector search Access: Free
Parameters* required
limitnumberMaximum number of results to return.
project_idstringProject ID (UUID).
querystringSearch query string.
workspace_idstringWorkspace ID (UUID).
search_hybridHybrid search (semantic + keyword) Access: Free4 params
Hybrid search (semantic + keyword) Access: Free
Parameters* required
limitnumberMaximum number of results to return.
project_idstringProject ID (UUID).
querystringSearch query string.
workspace_idstringWorkspace ID (UUID).
search_keywordKeyword search Access: Free4 params
Keyword search Access: Free
Parameters* required
limitnumberMaximum number of results to return.
project_idstringProject ID (UUID).
querystringSearch query string.
workspace_idstringWorkspace ID (UUID).
memory_create_eventCreate a memory event for a workspace/project Access: Free6 params
Create a memory event for a workspace/project Access: Free
Parameters* required
contentstringFull content/body.
event_typestringInput parameter: event type.
metadataobjectInput parameter: metadata.
project_idstringProject ID (UUID).
titlestringShort descriptive title.
workspace_idstringWorkspace ID (UUID).
memory_list_eventsList memory events (optionally scoped) Access: Free3 params
List memory events (optionally scoped) Access: Free
Parameters* required
limitnumberMaximum number of results to return.
project_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
memory_create_nodeCreate a knowledge node with optional relations Access: Free6 params
Create a knowledge node with optional relations Access: Free
Parameters* required
contentstringFull content/body.
node_typestringInput parameter: node type.
project_idstringProject ID (UUID).
relationsarrayInput parameter: relations.
titlestringShort descriptive title.
workspace_idstringWorkspace ID (UUID).
memory_list_nodesList knowledge graph nodes Access: Free3 params
List knowledge graph nodes Access: Free
Parameters* required
limitnumberMaximum number of results to return.
project_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
memory_searchSearch memory events/notes Access: Free4 params
Search memory events/notes Access: Free
Parameters* required
limitnumberMaximum number of results to return.
project_idstringProject ID (UUID).
querystringSearch query string.
workspace_idstringWorkspace ID (UUID).
memory_decisionsList decision summaries from workspace memory Access: Free4 params
List decision summaries from workspace memory Access: Free
Parameters* required
categorystringOptional category filter. If not specified, returns all decisions regardless of category.
limitnumberMaximum number of results to return.
project_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
graph_relatedFind related nodes in the knowledge graph Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)4 params
Find related nodes in the knowledge graph Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)
Parameters* required
limitnumberMaximum number of results to return.
node_idstringNode ID (UUID).
project_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
graph_pathFind path between two nodes Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)4 params
Find path between two nodes Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)
Parameters* required
project_idstringProject ID (UUID).
source_idstringID for the source.
target_idstringID for the target.
workspace_idstringWorkspace ID (UUID).
graph_decisionsDecision history in the knowledge graph Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)3 params
Decision history in the knowledge graph Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)
Parameters* required
limitnumberMaximum number of results to return.
project_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
graph_dependenciesDependency graph query Access: Pro (Graph-Lite) (upgrade: https://contextstream.io/pricing)3 params
Dependency graph query Access: Pro (Graph-Lite) (upgrade: https://contextstream.io/pricing)
Parameters* required
include_transitivebooleanInclude transitive dependencies.
max_depthnumberMaximum traversal depth.
targetobjectInput parameter: target.
graph_call_pathFind call path between two targets Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)3 params
Find call path between two targets Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)
Parameters* required
max_depthnumberMaximum traversal depth.
sourceobjectInput parameter: source.
targetobjectInput parameter: target.
graph_impactAnalyze impact of a target node Access: Pro (Graph-Lite) (upgrade: https://contextstream.io/pricing)2 params
Analyze impact of a target node Access: Pro (Graph-Lite) (upgrade: https://contextstream.io/pricing)
Parameters* required
max_depthnumberMaximum traversal depth.
targetobjectInput parameter: target.
graph_ingestBuild and persist the dependency graph for a project. Runs async by default (wait=false) and can take a few minutes for larger repos. Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)2 params
Build and persist the dependency graph for a project. Runs async by default (wait=false) and can take a few minutes for larger repos. Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)
Parameters* required
project_idstringProject ID (UUID).
waitbooleanIf true, wait for ingestion to finish before returning. Defaults to false (async).
projects_getGet project details by ID Access: Free1 params
Get project details by ID Access: Free
Parameters* required
project_idstringProject ID (UUID).
projects_overviewGet project overview with summary information Access: Free1 params
Get project overview with summary information Access: Free
Parameters* required
project_idstringProject ID (UUID).
projects_statisticsGet project statistics (files, lines, complexity) Access: Free1 params
Get project statistics (files, lines, complexity) Access: Free
Parameters* required
project_idstringProject ID (UUID).
projects_filesList all indexed files in a project Access: Free1 params
List all indexed files in a project Access: Free
Parameters* required
project_idstringProject ID (UUID).
projects_index_statusGet project indexing status Access: Free1 params
Get project indexing status Access: Free
Parameters* required
project_idstringProject ID (UUID).
projects_ingest_localRead ALL files from a local directory and ingest them for indexing. This indexes your entire project by reading files in batches. Automatically detects code files and skips ignored directories like node_modules, target, dist, etc. Access: Free4 params
Read ALL files from a local directory and ingest them for indexing. This indexes your entire project by reading files in batches. Automatically detects code files and skips ignored directories like node_modules, target, dist, etc. Access: Free
Parameters* required
overwritebooleanAllow overwriting existing files when write_to_disk is enabled
pathstringLocal directory path to read files from
project_idstringProject to ingest files into (defaults to current session project)
write_to_diskbooleanWhen true, write files to disk under QA_FILE_WRITE_ROOT before indexing (for testing/QA)
workspaces_getGet workspace details by ID Access: Free1 params
Get workspace details by ID Access: Free
Parameters* required
workspace_idstringWorkspace ID (UUID).
memory_get_eventGet a specific memory event by ID with FULL content (not truncated). Use this when you need the complete content of a memory event, not just the preview returned by search/recall. Access: Free1 params
Get a specific memory event by ID with FULL content (not truncated). Use this when you need the complete content of a memory event, not just the preview returned by search/recall. Access: Free
Parameters* required
event_idstringThe UUID of the memory event to retrieve
memory_update_eventUpdate a memory event Access: Free4 params
Update a memory event Access: Free
Parameters* required
contentstringFull content/body.
event_idstringEvent ID (UUID).
metadataobjectInput parameter: metadata.
titlestringShort descriptive title.
memory_delete_eventDelete a memory event Access: Free1 params
Delete a memory event Access: Free
Parameters* required
event_idstringEvent ID (UUID).
memory_timelineGet chronological timeline of memory events for a workspace Access: Free1 params
Get chronological timeline of memory events for a workspace Access: Free
Parameters* required
workspace_idstringWorkspace ID (UUID).
memory_summaryGet condensed summary of workspace memory Access: Free1 params
Get condensed summary of workspace memory Access: Free
Parameters* required
workspace_idstringWorkspace ID (UUID).
graph_circular_dependenciesDetect circular dependencies in project code Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)1 params
Detect circular dependencies in project code Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)
Parameters* required
project_idstringProject ID (UUID).
graph_unused_codeDetect unused code in project Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)1 params
Detect unused code in project Access: Elite/Team (Full Graph) (upgrade: https://contextstream.io/pricing)
Parameters* required
project_idstringProject ID (UUID).
session_initInitialize a new conversation session and automatically retrieve relevant context. This is the FIRST tool AI assistants should call when starting a conversation. Returns: workspace info, project info, recent memory, recent decisions, relevant context, and high-priority lessons...9 params
Initialize a new conversation session and automatically retrieve relevant context. This is the FIRST tool AI assistants should call when starting a conversation. Returns: workspace info, project info, recent memory, recent decisions, relevant context, and high-priority lessons...
Parameters* required
allow_no_workspacebooleanIf true, allow session_init to return connected even if no workspace is resolved (workspace-level tools may not work).
auto_indexbooleanAutomatically create and index project from IDE workspace (default: true)
context_hintstringRECOMMENDED: Pass the user's first message here for semantic search. This finds relevant context from ANY time, not just recent items.
folder_pathstringCurrent workspace/project folder path (absolute). Use this when IDE roots are not available.
include_decisionsbooleanInclude recent decisions (default: true)
include_recent_memorybooleanInclude recent memory events (default: true)
project_idstringProject to initialize context for
session_idstringCustom session ID (auto-generated if not provided)
workspace_idstringWorkspace to initialize context for
session_toolsGet an ultra-compact list of all available ContextStream MCP tools. Use this when you need to know what tools are available without reading full descriptions. Returns a token-efficient tool catalog (~120 tokens) organized by category. Format options: - 'grouped' (default): Cat...2 params
Get an ultra-compact list of all available ContextStream MCP tools. Use this when you need to know what tools are available without reading full descriptions. Returns a token-efficient tool catalog (~120 tokens) organized by category. Format options: - 'grouped' (default): Cat...
Parameters* required
categorystringFilter to specific category: Session, Search, Memory, Knowledge, Graph, Workspace, Project, AI
formatstringOutput format: grouped (default, ~120 tokens), minimal (~80 tokens), or full (~200 tokens)one of
grouped · minimal · fulldefault: groupedsession_get_user_contextRetrieve user preferences, coding style, and persona from memory. Use this to understand how the user likes to work and adapt your responses accordingly. Access: Free1 params
Retrieve user preferences, coding style, and persona from memory. Use this to understand how the user likes to work and adapt your responses accordingly. Access: Free
Parameters* required
workspace_idstringWorkspace to get user context from
workspace_associateAssociate a folder/repo with a workspace after user selection. Call this after session_init returns status='requires_workspace_selection' and the user has chosen a workspace. This persists the selection to .contextstream/config.json so future sessions auto-connect. Optionally...5 params
Associate a folder/repo with a workspace after user selection. Call this after session_init returns status='requires_workspace_selection' and the user has chosen a workspace. This persists the selection to .contextstream/config.json so future sessions auto-connect. Optionally...
Parameters* required
create_parent_mappingbooleanAlso create a parent folder mapping (e.g., /dev/maker/* -> workspace)
folder_pathstringAbsolute path to the folder/repo to associate
generate_editor_rulesbooleanGenerate AI editor rules for Windsurf, Cursor, Cline, Kilo Code, Roo Code, Claude Code, and Aider
workspace_idstringWorkspace ID to associate with
workspace_namestringWorkspace name for reference
workspace_bootstrapCreate a new workspace (user-provided name) and onboard the current folder as a project. This is useful when session_init returns status='requires_workspace_name' (no workspaces exist yet) or when you want to create a new workspace for a repo. Behavior: - Creates a workspace w...8 params
Create a new workspace (user-provided name) and onboard the current folder as a project. This is useful when session_init returns status='requires_workspace_name' (no workspaces exist yet) or when you want to create a new workspace for a repo. Behavior: - Creates a workspace w...
Parameters* required
auto_indexbooleanAutomatically create and index project from folder (default: true)
context_hintstringOptional context hint for session initialization
create_parent_mappingbooleanAlso create a parent folder mapping (e.g., /dev/company/* -> workspace)
descriptionstringOptional workspace description
folder_pathstringAbsolute folder path (defaults to IDE root/cwd)
generate_editor_rulesbooleanGenerate AI editor rules in the folder for automatic ContextStream usage
visibilitystringWorkspace visibility (default: private)one of
private · publicworkspace_namestringName for the new workspace (ask the user)
session_captureAutomatically capture and store important context from the conversation. Use this to persist decisions, insights, preferences, or important information. Access: Free8 params
Automatically capture and store important context from the conversation. Use this to persist decisions, insights, preferences, or important information. Access: Free
Parameters* required
contentstringFull content/details to capture
event_typestringType of context being capturedone of
conversation · decision · insight · preference · task · bugimportancestringImportance levelone of
low · medium · high · criticalproject_idstringProject ID (UUID).
session_idstringSession ID to associate with this capture
tagsarrayTags for categorization
titlestringBrief title for the captured context
workspace_idstringWorkspace ID (UUID).
session_capture_lessonCapture a lesson learned from a mistake or correction. Use this when the user corrects you, expresses frustration, or points out an error. These lessons are surfaced in future sessions to prevent repeating the same mistakes. Example triggers: - User says "No, you should..." or...9 params
Capture a lesson learned from a mistake or correction. Use this when the user corrects you, expresses frustration, or points out an error. These lessons are surfaced in future sessions to prevent repeating the same mistakes. Example triggers: - User says "No, you should..." or...
Parameters* required
categorystringCategory of the lessonone of
workflow · code_quality · verification · communication · project_specificimpactstringWhat went wrong (e.g., "Production 404 errors - broken landing page")
keywordsarrayKeywords for matching in future contexts (e.g., ["git", "images", "assets", "push"])
preventionstringHow to prevent in future (e.g., "Run git status to check untracked files before pushing")
project_idstringProject ID (UUID).
severitystringSeverity: critical for production issues, high for breaking changes, medium for workflow, low for minorone of
low · medium · high · criticaldefault: mediumtitlestringLesson title - what to remember (e.g., "Always verify assets in git before pushing")
triggerstringWhat action caused the problem (e.g., "Pushed code referencing images without committing them")
workspace_idstringWorkspace ID (UUID).
session_get_lessonsRetrieve lessons learned from past mistakes and corrections. Use this to check for relevant warnings before taking actions that have caused problems before. Returns lessons filtered by: - Query: semantic search for relevant lessons - Category: workflow, code_quality, verificat...6 params
Retrieve lessons learned from past mistakes and corrections. Use this to check for relevant warnings before taking actions that have caused problems before. Returns lessons filtered by: - Query: semantic search for relevant lessons - Category: workflow, code_quality, verificat...
Parameters* required
categorystringFilter by categoryone of
workflow · code_quality · verification · communication · project_specificlimitnumberMaximum lessons to returndefault: 10
project_idstringProject ID (UUID).
querystringSearch for relevant lessons (e.g., "git push images")
severitystringFilter by minimum severityone of
low · medium · high · criticalworkspace_idstringWorkspace ID (UUID).
session_smart_searchSearch memory with automatic context enrichment. Returns memory matches, relevant code, and related decisions in one call. Access: Free5 params
Search memory with automatic context enrichment. Returns memory matches, relevant code, and related decisions in one call. Access: Free
Parameters* required
include_decisionsbooleanInclude related decisions (default: true)
include_relatedbooleanInclude related context (default: true)
project_idstringProject ID (UUID).
querystringWhat to search for
workspace_idstringWorkspace ID (UUID).
session_rememberQuick way to store something in memory. Use natural language. Example: "Remember that I prefer TypeScript strict mode" or "Remember we decided to use PostgreSQL" Access: Free5 params
Quick way to store something in memory. Use natural language. Example: "Remember that I prefer TypeScript strict mode" or "Remember we decided to use PostgreSQL" Access: Free
Parameters* required
await_indexingbooleanIf true, wait for indexing to complete before returning. This ensures the content is immediately searchable.
contentstringWhat to remember (natural language)
importancestringInput parameter: importance.one of
low · medium · highproject_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
session_recallQuick way to recall relevant context. Use natural language. Example: "What were the auth decisions?" or "What are my TypeScript preferences?" Access: Free3 params
Quick way to recall relevant context. Use natural language. Example: "What were the auth decisions?" or "What are my TypeScript preferences?" Access: Free
Parameters* required
project_idstringProject ID (UUID).
querystringWhat to recall (natural language)
workspace_idstringWorkspace ID (UUID).
generate_editor_rulesGenerate AI rule files for editors (Windsurf, Cursor, Cline, Kilo Code, Roo Code, Claude Code, Aider). These rules instruct the AI to automatically use ContextStream for memory and context. Supported editors: codex, windsurf, cursor, cline, kilo, roo, claude, aider Access: Free8 params
Generate AI rule files for editors (Windsurf, Cursor, Cline, Kilo Code, Roo Code, Claude Code, Aider). These rules instruct the AI to automatically use ContextStream for memory and context. Supported editors: codex, windsurf, cursor, cline, kilo, roo, claude, aider Access: Free
Parameters* required
additional_rulesstringAdditional project-specific rules to append
dry_runbooleanIf true, return content without writing files
editorsarrayWhich editors to generate rules for. Defaults to all.
folder_pathstringAbsolute path to the project folder
modestringRule verbosity mode (default: minimal)one of
minimal · fullproject_namestringProject name to include in rules
workspace_idstringWorkspace ID to include in rules
workspace_namestringWorkspace name to include in rules
session_summaryGet a compact, token-efficient summary of workspace context (~500 tokens). This is designed to replace loading full chat history in AI prompts. Returns: workspace/project info, top decisions (titles only), preferences, memory count. Use this at conversation start instead of lo...3 params
Get a compact, token-efficient summary of workspace context (~500 tokens). This is designed to replace loading full chat history in AI prompts. Returns: workspace/project info, top decisions (titles only), preferences, memory count. Use this at conversation start instead of lo...
Parameters* required
max_tokensnumberMaximum tokens for summary (default: 500)
project_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
session_compressExtract and store key information from chat history as memory events. This allows clearing chat history while preserving important context. Use at conversation end or when context window is getting full. Extracts: - Decisions made - User preferences learned - Insights discover...4 params
Extract and store key information from chat history as memory events. This allows clearing chat history while preserving important context. Use at conversation end or when context window is getting full. Extracts: - Decisions made - User preferences learned - Insights discover...
Parameters* required
chat_historystringThe chat history to compress and extract from
extract_typesarrayTypes of information to extract (default: all)
project_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
session_deltaGet new context added since a specific timestamp. Useful for efficient context synchronization without reloading everything. Returns: - Count of new decisions and memory events - List of new items with titles and timestamps Use case: AI can track what's new since last session_...4 params
Get new context added since a specific timestamp. Useful for efficient context synchronization without reloading everything. Returns: - Count of new decisions and memory events - List of new items with titles and timestamps Use case: AI can track what's new since last session_...
Parameters* required
limitnumberMaximum items to return (default: 20)
project_idstringProject ID (UUID).
sincestringISO timestamp to get changes since (e.g., "2025-12-05T00:00:00Z")
workspace_idstringWorkspace ID (UUID).
context_smart**CALL THIS BEFORE EVERY AI RESPONSE** to get relevant context. This is the KEY tool for token-efficient AI interactions. It: 1. Analyzes the user's message to understand what context is needed 2. Retrieves only relevant context in a minified, token-efficient format 3. Replace...5 params
**CALL THIS BEFORE EVERY AI RESPONSE** to get relevant context. This is the KEY tool for token-efficient AI interactions. It: 1. Analyzes the user's message to understand what context is needed 2. Retrieves only relevant context in a minified, token-efficient format 3. Replace...
Parameters* required
formatstringContext format (default: minified)one of
minified · readable · structuredmax_tokensnumberMaximum tokens for context (default: 800)
project_idstringProject ID (UUID).
user_messagestringThe user message to analyze and get context for
workspace_idstringWorkspace ID (UUID).
reminders_listList all reminders for the current user. Returns: reminders with title, content, remind_at, priority, status, and keywords. Can filter by status (pending, completed, dismissed, snoozed) and priority (low, normal, high, urgent). Use this to see what reminders you have set. Acce...5 params
List all reminders for the current user. Returns: reminders with title, content, remind_at, priority, status, and keywords. Can filter by status (pending, completed, dismissed, snoozed) and priority (low, normal, high, urgent). Use this to see what reminders you have set. Acce...
Parameters* required
limitnumberMaximum reminders to return (default: 20)
prioritystringFilter by priorityone of
low · normal · high · urgentproject_idstringProject ID (UUID).
statusstringFilter by statusone of
pending · completed · dismissed · snoozedworkspace_idstringWorkspace ID (UUID).
reminders_activeGet active reminders that are pending, overdue, or due soon. Returns: reminders with urgency levels (overdue, due_soon, today, upcoming). Optionally provide context (e.g., current task description) to get contextually relevant reminders. Use this to see what reminders need att...4 params
Get active reminders that are pending, overdue, or due soon. Returns: reminders with urgency levels (overdue, due_soon, today, upcoming). Optionally provide context (e.g., current task description) to get contextually relevant reminders. Use this to see what reminders need att...
Parameters* required
contextstringOptional context to match relevant reminders (e.g., current task)
limitnumberMaximum reminders to return (default: 10)
project_idstringProject ID (UUID).
workspace_idstringWorkspace ID (UUID).
reminders_createCreate a new reminder for a specific date/time. Set reminders to be notified about tasks, follow-ups, or important dates. Priority levels: low, normal, high, urgent Recurrence: daily, weekly, monthly (optional) Example: Create a reminder to "Review PR #123" for tomorrow at 10a...8 params
Create a new reminder for a specific date/time. Set reminders to be notified about tasks, follow-ups, or important dates. Priority levels: low, normal, high, urgent Recurrence: daily, weekly, monthly (optional) Example: Create a reminder to "Review PR #123" for tomorrow at 10a...
Parameters* required
contentstringReminder details/description
keywordsarrayKeywords for contextual surfacing
prioritystringPriority level (default: normal)one of
low · normal · high · urgentproject_idstringProject ID (UUID).
recurrencestringRecurrence patternone of
daily · weekly · monthlyremind_atstringWhen to remind (ISO 8601 datetime, e.g., "2025-01-15T10:00:00Z")
titlestringReminder title (brief, descriptive)
workspace_idstringWorkspace ID (UUID).
reminders_snoozeSnooze a reminder until a later time. Use this to postpone a reminder without dismissing it. Common snooze durations: - 1 hour: add 1 hour to current time - 4 hours: add 4 hours - Tomorrow: next day at 9am - Next week: 7 days from now Access: Free2 params
Snooze a reminder until a later time. Use this to postpone a reminder without dismissing it. Common snooze durations: - 1 hour: add 1 hour to current time - 4 hours: add 4 hours - Tomorrow: next day at 9am - Next week: 7 days from now Access: Free
Parameters* required
reminder_idstringID of the reminder to snooze
untilstringWhen to resurface the reminder (ISO 8601 datetime)
reminders_completeMark a reminder as completed. Use this when the task or action associated with the reminder is done. Access: Free1 params
Mark a reminder as completed. Use this when the task or action associated with the reminder is done. Access: Free
Parameters* required
reminder_idstringID of the reminder to complete
reminders_dismissDismiss a reminder without completing it. Use this to remove a reminder that is no longer relevant. Access: Free1 params
Dismiss a reminder without completing it. Use this to remove a reminder that is no longer relevant. Access: Free
Parameters* required
reminder_idstringID of the reminder to dismiss
ContextStream MCP Server
Give every coding agent persistent memory, semantic code search, and the right project context on every turn.
Works with Claude Code, Cursor, VS Code + Copilot, Windsurf, Cline, Roo Code, Kilo Code, Codex, OpenCode, Aider, Antigravity, and other MCP clients.
Install in 30 seconds · Benchmarks · Documentation · Start free
Your AI should not have to rediscover the repository, re-learn the architecture, or repeat last week's mistake every time a chat starts. ContextStream gives MCP clients a shared context layer that remembers decisions and lessons, searches code by meaning, maps dependencies, and restores the thread after compaction.
The server in this repository is the canonical, MIT-licensed Rust
implementation. The recommended editor transport is the hosted MCP at
https://mcp.contextstream.io/mcp; a small local helper keeps explicitly linked
checkouts in sync.
Benchmarks: strong memory, strong retrieval, published methodology
We publish the full method, confidence intervals, comparison caveats, and known losses—not just the winning number. The product-level figures below are from the published July 2026 report; the LongMemEval-S result was published June 14, 2026.
| Benchmark | ContextStream result | What was measured |
|---|---|---|
| LongMemEval-S | 90.0% | 450/500 correct on the full suite; ~115k-token multi-session histories; official pinned GPT-4o judge; gpt-5.5 reader with self-consistency k=3 (89.6% single-shot); Wilson 95% CI [87.1%, 92.3%] |
| Pinned-repo code search | 99.38% Recall@10 | 93.83% Primary@10 and 95 ms successful-response p50 over 177 judged-valid queries, averaged across two cold and two warm runs; 0/708 failed requests; no local checkout |
| Agentic project memory | 96% task success | Up from 58% without the memory layer; one repository, one agent model, three trials per cell, so this result is directional |
On the standard memory benchmark, 90.0% is statistically above supermemory's published 85.4% and statistically tied with Zep's published 90.2%. Those comparisons use each vendor's published result; they are not presented as a private same-run head-to-head.
Read the full benchmark report, methodology, per-category results, and losses →
The repository also includes Criterion suites for JSON-RPC parsing, tool
dispatch, response serialization, and stdio/HTTP transport primitives. Run
them locally with cargo bench --locked -p mcp-server.
Install in 30 seconds
Pick the command for your platform. Each installer downloads the current native release and launches the guided setup wizard.
macOS and Linux
curl -fsSL https://contextstream.io/scripts/setup.sh | bash
Native releases are published for macOS Apple Silicon and Intel, plus Linux ARM64 and x64.
Windows PowerShell
irm https://contextstream.io/scripts/setup.ps1 | iex
The Windows installer uses the native x64 binary and installs it in your user profile, so it does not need an administrator shell.
npm alternative (Node.js 20+)
npx -y @contextstream/mcp-server@latest setup
The dependency-free npm launcher selects the matching release artifact, verifies its SHA-256 checksum, caches it by exact package version, and starts the Rust binary. It supports macOS ARM64/x64, Linux ARM64/x64, and Windows x64.
What the wizard does
The wizard signs you in, detects supported editors, writes their hosted MCP configuration and managed rules, installs lifecycle hooks where supported, links the exact project folder you choose, starts indexing, and checks the result. You review the choices before it makes changes.
After setup, restart the editor and try:
Use ContextStream to summarize this repository, cite the files you relied on, and tell me one active decision or risk that should affect my next change.
Want to inspect the changes first? Every setup write can be previewed:
contextstream-mcp setup --dry-run
Already configured and troubleshooting? Run:
contextstream-mcp doctor --scope=all --only-configured
No-install hosted connection
If your MCP client supports Streamable HTTP and OAuth, connect directly—there is no local MCP process to keep running:
{
"mcpServers": {
"contextstream": {
"type": "http",
"url": "https://mcp.contextstream.io/mcp"
}
}
}
Some clients use a different root key or config format, which is why the setup wizard is the easiest path. The hosted gateway is the default; the installed helper only syncs the local checkouts you explicitly link.
Clients that require stdio can use the verified npm launcher:
{
"mcpServers": {
"contextstream": {
"command": "npx",
"args": ["-y", "@contextstream/mcp-server@latest"],
"env": {
"CONTEXTSTREAM_API_KEY": "your-api-key"
}
}
}
}
For production automation, pin an exact 1.x package version instead of
latest.
What your agent gets
| Capability | What changes in practice |
|---|---|
| Grounded context | Each turn starts with the relevant workspace rules, prior decisions, lessons, and current project state. |
| Semantic code search | Find implementations by intent, with hybrid, keyword, pattern, exhaustive, and refactor-aware modes when precision matters. |
| Durable memory | Decisions, plans, tasks, runbooks, preferences, and session transcripts remain queryable across chats and tools. |
| Compaction survival | Important state is checkpointed before long conversations compact and restored afterward. |
| Dependency intelligence | Trace blast radius, dependencies, circular imports, complexity, and unused code without manually rebuilding the graph. |
| Agent handoffs | Package verified state and next steps into durable handoffs and portable ContextCapsules. |
| Team context | Bring shared workspace knowledge and supported GitHub, Slack, and Notion context into the same MCP surface. |
The default tool surface is grouped into focused domains—context, search,
memory, session, graph, entity, capsule, project, workspace,
skill, media, vcs, and more—so agents get broad capability without an
enormous tool-registration prompt.
Supported editors and agents
The wizard currently detects and configures:
- Claude Code
- Cursor
- Windsurf
- VS Code + GitHub Copilot
- Cline
- Kilo Code
- Roo Code
- OpenAI Codex
- Aider
- Antigravity
- OpenCode
Anything else that speaks the Model Context Protocol can connect to the hosted URL or use the stdio launcher.
What is open source?
This repository contains the canonical Rust MCP server, protocol types, client and session layers, editor setup, hooks, public transport code, release automation, and the MongoDB-free acceleration foundation. It builds as seven workspace crates:
mcp-servermcp-toolsmcp-clientmcp-sessionmcp-typesmcp-model-registrymcp-acceleration-products
The hosted backend, credentials, production deployment configuration, and
premium provider implementations are not included. Public-build compatibility
traits remain no-op so the core protocol surface stays stable. See
docs/architecture.md for the boundary.
Privacy and data controls
Setup shows a disclosure before changing editor configuration or indexing a
project. You choose the folder and workspace relationship; only matching source
files from a validated project can be indexed. Transcript capture, local Git
metadata capture, ignore rules, kill switches, retention behavior, and deletion
paths are documented in docs/data-handling.md.
Review that document before using setup --yes in automation. The CLI also
provides --dry-run for setup, repair, migration, and uninstall workflows.
Build and test from source
The workspace uses the Rust toolchain pinned in rust-toolchain.toml.
cargo build --locked -p mcp-server --bin contextstream-mcp
cargo test --locked --workspace
cargo bench --locked -p mcp-server
To build the hosted HTTP gateway foundation:
cargo build --locked --release -p mcp-server --bin contextstream-mcp \
--features remote-acceleration
Release integrity
Every release publishes native binaries, checksums.txt, version.json, and
an SPDX SBOM. The npm launcher downloads only from the immutable path for its
exact package version, verifies manifests and binary SHA-256, uses an atomic
versioned cache, and supports verified offline reuse with
CONTEXTSTREAM_MCP_OFFLINE=1.
See docs/release.md for the release contract and
the latest GitHub release
for artifacts.
All 1.x npm releases preserve these executable aliases:
mcp-servercontextstream-mcpcontextstream-hook(forwards to the Rusthooksubcommand)
Contributing and security
Issues and pull requests are welcome. See CONTRIBUTING.md,
SECURITY.md, and GOVERNANCE.md. Contributions
must include the Developer Certificate of Origin sign-off (git commit -s).
The code is MIT licensed. ContextStream trademarks and the hosted-service boundary are described in NOTICE.
Stop teaching your coding agent the same project twice.
Install ContextStream ·
Start free ·
Read the docs
Featured
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →Categories
Registryactive
Package@contextstream/mcp-server
TransportSTDIO
UpdatedJan 2, 2026
