Product Decisions This Supports
Note on fit: despite the framing, this is a Symfony bundle, not a Laravel package. There is no Laravel integration, no service provider, and no Telescope/Debugbar bridge. Treat it as prior art and a market signal, not a dependency candidate.
- Build vs. buy, answered by the vendor itself. The author has deprecated the package and redirected users to Symfony AI Mate. Any "should we adopt this?" decision is already made: no new adoption, and existing users should plan a migration. This is a rare case where the build-vs-buy question resolves to "the upstream framework now owns it."
- Validates a real product category: AI-native debugging. ~5.7k installs with only 11 stars and 0 dependents tells a specific story — a small, quiet, real user base that wanted AI coding agents to see runtime truth (requests, queries, timings) instead of guessing from source. That's a genuine unmet need worth putting on a roadmap, independent of this package's fate.
- Roadmap input for a first-party equivalent. For a Laravel shop, the analogous artifact is an MCP server over Laravel Telescope / Pulse / Debugbar data. This repo is a cheap reference architecture — JSON-RPC over MCP, read-only accessors onto profiler storage — that de-risks a 1–2 week internal spike.
- Prioritization signal on observability. Requests, queries, and timings are the three data classes the author chose to expose first. That's a useful, evidence-backed ranking of what agents actually need from a runtime, and it should inform the order of any internal observability API work.
- Ecosystem-timing lesson for planning. Shipped March 2025, ahead of any official Symfony solution — and was superseded within roughly a year. The PM takeaway: when building on a framework boundary that the framework vendor is actively investing in, budget for a short shelf life and design for replacement (thin adapter, no deep coupling).
When to Consider This Package
Consider it only as a time-boxed reference, not a dependency:
- You are on Symfony 6/7 and want a same-day, throwaway proof that "AI agent reads profiler data" changes debugging behavior — install as a
require-dev dependency, pinned, scoped to local development, then uninstall.
- You are designing an MCP observability tool for Laravel and want a concrete, readable implementation to copy patterns from (data shaping, read-only tool surface, transport wiring) rather than inventing the schema.
- You want a low-cost artifact to pressure-test an internal proposal before spending engineering cycles.
Look elsewhere — or don't adopt at all — when:
- You're on Laravel. No compatibility path exists; Laravel Telescope/Pulse plus a thin MCP wrapper is the only viable route.
- You need support, security review, or an SLA. Deprecated, single maintainer, 0 dependents, no vendor backing. This fails any standard third-party risk gate.
- The profiler contains production-shaped data. Profiler payloads routinely include SQL, request bodies, headers, session identifiers, and env-derived config. An MCP server granting an LLM read access to that is a data-governance decision, not a tooling decision — and it must never run against production.
- You want a maintained baseline. Go straight to Symfony AI Mate if you're on Symfony. Adopting the deprecated predecessor means paying a migration cost for no interim benefit.
- You're evaluating on popularity or ecosystem gravity. 11 stars and 0 dependents is below any reasonable threshold for a default choice; the 5.7k installs are interesting as demand evidence, not as a quality signal.
How to Pitch It (Stakeholders)
To executives (one paragraph): A small open-source tool appeared in March 2025 that let AI coding assistants read application runtime data — the actual requests, database queries, and timings — instead of reasoning only from source code. It attracted a modest but real user base and was then deprecated by its own author in favor of an official Symfony solution. We are not recommending adoption. We are recommending we read it as market evidence: AI-assisted debugging is becoming a genuine workflow, framework vendors are moving into the space, and the window to build our own version on Laravel is best spent now, as a small internal tool, rather than bought. Total exposure if we do nothing is low; total upside if we build it is faster diagnosis of performance and query problems.
To engineering (short): This is a Symfony-only MCP server exposing profiler data to AI agents. It's deprecated, unmaintained, and superseded — do not add it to any project. What we want from it is the pattern: a read-only MCP surface over Telescope (requests, queries, timings), dev-environment-only, no production credentials, no PII in tool output. Propose a one-to-two-week spike to build a Laravel equivalent, using this repo as the reference implementation and Symfony AI Mate as the design comparison. Flag it as exploratory and non-production until we've reviewed what data the agent can actually see.