Enables MCP-compatible clients and autonomous agents to dynamically compress long-horizon context by pruning verbose tool outputs, HTML boilerplate, and semantic redundancy by 40-70%. It reduces token burn and latency through deterministic, zero-dependency processing.
MCP servers
Compare source listings, connection options and authentication requirements.
Showing 3,769–3,792 of 165,916 resultsPage 158 of 6,914
Candidate groups, not verified unique MCPs. How grouping works
Enables AI agents and MCP clients to dynamically select few-shot in-context examples using semantic token overlap, cosine similarity, and MMR diversity ranking to optimize prompts deterministically without external dependencies.
Enables MCP-compliant clients and autonomous agents to pick optimal few-shot in-context exemplars by combining semantic token overlap, cosine similarity, and Maximal Marginal Relevance diversity ranking for prompt optimization. Runs deterministically on the Python standard library with zero external dependencies, avoiding frontier LLM round-trips for sub-millisecond selection.
Enables autonomous agents to snapshot execution-tree and context states, evaluate branch viability, and roll back to previous checkpoints when an exploration path fails without restarting from scratch.
Enables autonomous agents and MCP clients to decompose text into atomic claims and verify them against reference knowledge, detecting ungrounded claims, numerical contradictions, and hallucinated entities.
Enables agents to maintain multi-tier memory across working scratchpad, dialogue buffer, and episodic stores with recency decay. Supports hybrid BM25/vector retrieval, rank fusion, graph expansion, and context reordering for optimized MCP-based memory and RAG workflows.
Enables autonomous agents and MCP-compatible clients to enforce policy-approved actions using risk tiers, reversible checkpoints, and token approval bounds.
Enables MCP-compatible agents and IDEs to retrieve episodic long-term memories using hybrid BM25 keyword search, semantic n-gram cosine similarity, and recency decay ranking, with deterministic zero-dependency execution.
Enables MCP-compatible agents and IDEs to deterministically extract JSON and auto-repair malformed schemas, including truncated braces, markdown wrappers, and invalid types, without LLM round-trips.
Enables long-horizon conversational memory compaction and episodic anchor extraction for AI agents, providing deterministic memory management through a zero-dependency MCP server.
Enables AI agents and MCP clients to heuristically parse multi-column document dumps, aligned spacing, and CSV/TSV input, converting them into clean GitHub Flavored Markdown tables and typed JSON. Runs fully offline on the Python standard library with zero external dependencies, returning deterministic results without LLM round-trips.
Enables AI agents and MCP clients to convert ASCII/OCR text dumps, aligned spacing, and CSV/TSV data into clean GitHub Flavored Markdown tables and typed JSON deterministically.
Enables autonomous AI agents and MCP-compliant clients to detect and block adversarial prompt injections, jailbreak attempts, system prompt exfiltration, and tool hijacking via a zero-dependency, sub-millisecond heuristic firewall.
Enables AI agents and multi-agent frameworks to throttle outbound LLM API calls with multi-provider token-bucket and sliding-window limiting, plus predictive token-consumption budgeting to avoid TPM/RPM throttling. It runs deterministically over MCP with zero dependencies, returning rate-limit decisions and telemetry for any client such as Claude Desktop or Cursor.
Enables MCP clients to run deterministic, zero-dependency recursive self-critique and reflection loops over a payload and optional options, iteratively evaluating and revising agent output to autonomously repair logic defects, test failures, and hallucinations. Supports direct JSON-RPC stdio integration into MCP-compliant clients such as Claude Desktop and Cursor.
Enables autonomous agents to detect configuration drift, corrupted execution state, and missing runtime resources, then perform root-cause classification and automated snapshot rollback for self-healing runtime recovery.
Enables agents and MCP-compliant clients to run multi-language code in a deterministic, zero-dependency sandbox where AST security analysis blocks dangerous syscalls and isolated execution telemetry is captured. It exposes this through JSON-RPC 2.0 stdio, letting frameworks like LangGraph, CrewAI, Cursor, and Claude Desktop validate and execute agent-generated code safely.
Enables agents and MCP clients to safely execute and verify multi-language code in a deterministic sandbox, using AST security analysis to block dangerous system calls and capture isolated execution telemetry.
Enables aggregating divergent LLM reasoning trajectories via multi-path chain-of-thought self-consistency consensus scoring and majority voting into high-confidence deterministic outputs. Plugs into any MCP-compliant client over JSON-RPC 2.0 stdio with zero external dependencies.
Enables high-throughput semantic query caching and approximate deduplication by combining exact SHA-256 matching, cosine similarity threshold checks, and LRU eviction.
Enables AI agents and developers to dynamically tokenize session context windows and compact dialogue history into structured outputs via MCP, CLI, or Python client. Runs on pure standard library Python with no external dependencies for deterministic, low-latency execution.
Enables MCP-compatible clients to run speculative decoding draft-model verification, calculating acceptance probabilities and returning structured execution telemetry with zero external dependencies.
Enables agents to orchestrate speculative decoding runs and deterministically verify draft-model acceptance probabilities via MCP.
Enables autonomous web agents to prune DOM trees, canonicalize URLs against anti-crawler traps, auto-map form input schemas, and convert HTML/Markdown tables into typed JSON records. It also scores agent execution trajectories against gold benchmarks using action-level precision, recall, and Levenshtein sequence edit distance.
What this catalog covers
- Source listings
- 167,896
- Candidate groups
- 168,272
- Setup-eligible endpoints
- 30,743
Sources observed 2026-10-24–2026-11-11. Listings can overlap across sources. Groups are identity candidates; setup eligibility is not a successful connection test.
SourcesMCP data from Glama connectorsMCP data from Glama source codeOfficial MCP RegistryClaude connector directoryAPI Tracker · publisher documentationSlack MCP directory
Default browsing excludes 2,356 specialized utility groups across the catalog. They remain available through keyword search and direct links. This editorial filter does not change the indexed totals or establish publisher quality.
Categories currently cover records matched to our reviewed catalog; other listings remain Uncategorized. Source-only projects have no generated connection option. Browse reviewed official MCPs →
- Glama connectors
API result set; pagination may cap at approximately 1,000 connectors. Not exhaustive.
- Glama source code
Complete chronological source-code catalog response; includes source-reported author classifications, not publisher attestations or ready-to-use endpoints.
- Official MCP Registry
Complete paginated registry response; registry membership is not publisher verification.
- Claude connector directory
Complete public Claude directory MCP components. Source tiers, tools, icons and sign-in hints are declarations, not publisher or runtime verification.
- API Tracker · publisher documentation
Reviewed publisher and third-party remote, local and setup-only MCP listings. Publisher status is retained per exact connection. Discoverability does not establish connection readiness.
- Slack MCP directory
Reviewed additions discovered through the public Slack Marketplace MCP category. Complete discovery inventory is retained separately; membership does not prove external-client compatibility.
Listings are grouped by exact public endpoints, exact package sets, or explicit reviewed identity bindings. Shared names or repositories alone do not merge implementations. Official includes publisher evidence and accepted source author classifications. It does not mean independently verified, runtime-tested or security-certified.