context-mode Platform
observability layer for AI coding agents
Last verified:
What is context-mode Platform?
context-mode Platform is an enterprise AI engineering observability platform that measures the engineering signal AI tools like Claude Code, Cursor, Copilot, Codex, and Gemini are actually shipping. It provides role-aware observability across 14 AI assistants, evaluating 222 patterns across 9 categories including adoption, effectiveness, quality, cost, and security. The platform sits on top of the open-source Context Mode plugin that 250,000+ developers already trust locally.
The tool ingests structured session events (tool calls, file paths touched, error counts, token usage, commit messages) without ever collecting prompt content or source code. It features 13 first-class MCP tools for agentic access, a web dashboard, REST API for CI/CD integration, and supports 7 personas: IC, Manager, EM, CTO, Board, Security, and FinOps. Key capabilities include finding blockers, tracking commit quality, measuring ROI, identifying security risks, and analyzing cost efficiency.
context-mode Platform is designed for engineering leaders (CTOs, Engineering Managers), Individual Contributors wanting personal productivity insights, Security teams (CISO/AppSec), FinOps professionals tracking AI spend, and DevOps leads monitoring release quality. Organizations like Microsoft, Google, Meta, ByteDance, GitHub, Stripe, and Anthropic use the open-source plugin locally.
context-mode Platform pricing
Pricing model: Freemium
Pro tier: $20 per seat per month, billed monthly. Per-seat means per member of your org (active or invited). No annual variant, no volume discount, no founder discount, no grandfathering, no Enterprise SKU, no free tier, no trial. Self-serve at platform.context-mode.com. Cancel any time via Stripe Customer Portal - no refunds for current period. Member adds/removes sync pro-rated on next invoice. Example: 47 engineers = $940/month.
context-mode Platform pros
- 98% context window reduction through sandboxed tool output
- 222 empirically-derived patterns across 9 categories
- Supports 14 AI assistants including Claude Code, Cursor, Copilot, Codex, Gemini
- Privacy-first design - never collects prompt content or source code
- 13 first-class MCP tools for agentic access
- Role-aware RBAC enforced at schema layer (Owner/Manager/Member)
- 7 persona-specific views for CTO, EM, IC, CISO, FinOps, DevOps
- 90-day data retention included
- Real-time blocker detection for engineering managers
- Commit quality classification with Conv Commits prefix detection
- Auto-redacted secret leak detection for security teams
- Cost per commit tracking for FinOps unit economics
- Integrations with Slack, Linear, Jira, GitHub Actions, Notion, Teams, Discord
- Open-source plugin free forever with 250K+ developers
- Sub-second p95 performance for org-rollups
context-mode Platform cons
- No free tier for the Platform/Insight layer
- No trial period available
- $20/seat/month with no volume discounts
- No annual plan option
- No refunds policy
- SSO/SAML not included in Pro tier
- SOC 2 compliance requires self-hosted OSS version
- Member scope cannot view org-level data
- Data forwarding only works with explicit token opt-in
- SQLite database is ephemeral and session-scoped for OSS plugin
Frequently asked questions about context-mode Platform
How much does Insight cost?
Pro tier is $20 per seat per month. Per-seat means per member of your org (active or invited). Billed monthly. No annual variant, no volume discount, no Enterprise SKU, no free tier, no trial. Self-serve at platform.context-mode.com.
Can I try it free?
The context-mode OSS collector is free forever - install it from the GitHub README to run insight locally on your own machine. The Platform layer (platform.context-mode.com) is paid from day one. There is no Insight trial.
What data do you ingest?
Structured session events emitted by the OSS plugin's hook system - never prompt content, never source code, never file content. The envelope covers: tool calls (name, duration, exit status), prompt boundaries, file paths touched, error and retry counts, commit message subjects, agent token usage with derived cost, session metadata, ExitPlanMode signals, ExitWorktree signals, SessionStart settings, MCP-server snapshots, Layer-3 streaming word-token counts, and 10 universal Unicode-property prompt features.
Does the OSS plugin still work without Insight?
Yes - and that is the default posture. The OSS plugin is licensed under Elastic License 2.0 and operates at the MCP protocol layer: raw data stays in a sandboxed subprocess and never enters the LLM context window. Nothing leaves the machine - no telemetry, no cloud sync, no usage tracking, no account required. Insight is an explicit opt-in: generate a token in the dashboard and the same plugin starts forwarding structured events. Remove the token, forwarding stops.
Who sees what?
Owner sees the org. Manager sees the teams listed in their managed_teams column. Member sees themselves plus anonymised peer benchmarks. Enforced at the MCP schema layer - a Member token literally cannot pass scope: org because the enum lacks the value.
What is your refund policy?
No refunds. You can cancel any time from the Stripe Customer Portal - your subscription ends at the close of the current billing period. No partial-month credits. Member adds and removes mid-cycle sync to Stripe as quantity changes, billed pro-rated on the next invoice.
Do you offer SSO/SAML/SOC 2?
Not in the Pro tier. The OSS collector is source-available under ELv2 and can be run self-hosted if your buying process requires those controls.
What is the agentic/MCP angle?
The remote MCP endpoint is a first-class surface - every dashboard widget has an MCP twin. The agent server ships 13 tools today: find_blockers, what_changed, top_patterns, how_is_team, compare_teams, how_am_i, identify_risk, engagement_health, top_members, activity_timeline, activity_breakdown, member_detail, and commit_quality_team. Tool schemas are dynamically computed per token - the schema itself enforces access boundaries.
How are the patterns built?
Empirically, from real session telemetry across the OSS adapter surface (16 platform integrations on OSS side; 14 currently exposed in Insight install CLI). The catalog ships 222 patterns across 9 categories (adoption, effectiveness, efficiency, quality, workflow, engagement, prompt, security, cost), each one wired through getMetricValue + AggregateData in the engine and independently testable. Every pattern is SQL + threshold + severity + suggested action, audited against existing event flows.
How does Insight relate to Context Mode Platform?
Context Mode Platform is the umbrella - the OSS plugin, the event ingestion layer, the pattern engine, the dashboard, the remote MCP. Insight is the first Solution built on top of it: the org-layer insight surface, role-tuned, persona-aware. Two more Solutions are on the roadmap: Audit turns the event stream into compliance reports for the CISO persona; Risk Score scores AI-authored changes inside a PR for the DevOps Lead persona, exposed as a GitHub Action.