Figma Context MCP
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
Description from the maker (original English)
Governed design-to-print authority for agents — signed provenance from model to machine, safety
These three are never merged into one score: opinions, measurements and popularity are different things.
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
Pick the AI app you use, then copy the settings.
The maker has not listed install details in the registry. See the maker’s README below.
Runs on your computer; the tool list appears in your AI app after you install it.
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 6, 2026
The registry name was verified with the GitHub account CYAN-ECLIPSE-TECHNOLOGIES-LLC, which owns the repo
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
This one runs on your computer; there is no internet endpoint
Runs on your computer and uses local settings instead
Runs on your computer; we can only read tool lists from internet endpoints
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).
The GitHub repo last changed –, and the latest version (0.1.0) reached the registry 11 days ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code.
In the same category: Figma Context MCP, Claude Real Video, Figma MCP Server
Servers in the same category from other makers, sorted by stars.
Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
Let any LLM watch a video locally — and search everything it has ever watched.
Figma’s official server gives AI your design data so it can turn designs into code or describe them accurately. Connects to Figma’s cloud.
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
*governed design-to-print authority for agents — signed provenance from model to machine, safety envelopes at the bridge, evidence for every print.*
Status: CERTIFIED(simulator) — Eclipse Bazaar Test Bench 2026-10-25, scoped re-bench r3 (standing card `../test-bench/results/eclipse-strata-mcp-2026-10-25-r3.json`, 12/12 scoped checks). Bench history, all cards retained append-only: bench 1 FAILED (35 of 45), scoped re-bench r2 FAILED (25 of 26 — D2 signer-anchor defect open), r3 CERTIFIED(simulator) with D2 closed. Simulator scope only — no live printer was exercised; no claim broader than CERTIFIED(simulator) is made.
Spec: ../eclipse-strata-spec-20261025.md (Silas Forge W-002, 2026-10-25). Renamed Eclipse Loft — owner decision 2026-10-25: built and certified as Eclipse Strata; the Stratasys trademark-screen risk was rejected and fallback 2 (Loft) adopted over Sinter (D7). Test Bench scorecards retain the Strata name as history.
Fabrication authority exercised only inside a signed mandate pinning agent × printer serial × design pin × slicer build × profile hash × material × thermal/volume/duration/budget envelope × site class × window. The bridge re-slices the pinned inputs at dispatch and canonical-hash-matches (§4.3: comments + volatile header stripped, rest byte-exact) before any printer contact. Every print and every refusal seals into a Blackbox-form evidence pack (16+ fields, §12) verifiable offline by Loft Proof.
server.py, stdio JSON-RPC, stdlib only) — exactly 11 tools: design_register, design_inspect, design_pin_get, printers_discover, session_open, job_prepare, job_dispatch, job_status, job_abort, telemetry_read, session_close. No raw G-code, axis, heater, config-write, code-exec, or pin-activation tool exists on the surface; forbidden names refuse -32602 with the named refusal (§11).strata_engine.py) — simulators only at P0: OctoPrint virtual printer + virtual Klipper/Moonraker (+ a virtual multi-function unit for non-print-head refusal). Slicer is a deterministic simulated PrusaSlicer/Orca behind the production interface (real CLI adapters slot in at P1+); re-derivation is real: identical pinned inputs re-derive the identical canonical hash, any content edit changes it.bridge_admin.py) — operator-only, bridge host, never MCP-reachable: toolchain registry, printer declaration, human-principal pin activation, pack export; admin acts receipted to data/bridge-admin.log.proof.py, tools/strata_proof.py) — offline pack verifier: PRINT_VERIFIED | PRINT_UNVERIFIED, failing element named, toolpath re-derived from the pack's pinned inputs. Verification requires a trusted signer anchor supplied to the verifier (verify_pack(..., registry=… / expected_fingerprint=…); CLI --registry FILE / --expected-fingerprint FP): the pack's signer fingerprint + embedded key are cross-checked against that local registry (§7.1) — unknown signers, registry key mismatches, and no-anchor presentations refuse naming UNVERIFIED_SIGNER.gauge.py, tools/strata_gauge.py) — static mandate+profile conformance: ENVELOPE_SOUND | ENVELOPE_UNSOUND, violations named.CONFIG_DRIFT refusal on firmware attestation change; config writes = permanent refusal (no T3 verb exists); T2 raw G-code refused; dead-man = telemetry loss (mandate gap threshold, default 30 s, 10–300) → state UNKNOWN + pause, resume only after re-attestation; mandate expiry mid-print ≠ kill (completes under monitoring unless hard_stop_at_expiry signed); PENDING_PHYSICAL_CONFIRM is honest waiting, never a bypass; readiness facts are labeled machine-attested / human-declared / unchecked, never upgraded. Registration split (§4.4/S6): design_register only PROPOSES; activation requires a human-principal signed pin record — agent self-pin refuses AGENT_SELF_PIN_REFUSED.
PinSigner port with local Ed25519 signer fallback (pure-stdlib; Herald upgrades verifiability, never behavior); Blackbox-form packs native; Tripwire-pattern ceilings native; Friction auto-file hook on adapter telemetry findings; Meter-shaped consumption records emitted read-side per print; Till = P3-gated port only (present, disabled, no active code paths); Span delegation slot present in the mandate grammar from P0 (inert); Census rejected (no port, no discovery targeting).
python3 -m unittest discover -s tests — 64 tests: happy-path full governed print on both simulators + pack verification, guards (coerce_int_param, parse_time, ref guard), canonical-hash volatile-invariance, and the §14 hostile-agent suite H1–H18 in bench shape (same fixtures, same expected refusal codes). Note: pure-Python Ed25519 makes the suite slow (~95 s); signing is the only non-trivial cost. python3 smoke.py — wire-level smoke (11 tools, discover, forbidden-name refusal).
Simulators only — no real printer was touched, no heat, no network calls, no vendor clouds, no Bambu path (Developer-Mode gate is P2+). A mandate layer never substitutes for a site's own risk assessment, fire precautions, or insurer terms (§6.6). On an open desktop, bridge custody is a discipline, not physics (§3). Geometry at P0 is declared geometry only (fix round D7): design_inspect returns the caller-declared bounding box from the parameter set plus arithmetic over it — no kernel/mesh inspection of the artifact is performed, and manifold / wall-thickness / triangle facts are reported unchecked (null), never echoed as verified. Packs are key-sealed (fix round D2): every pack carries an Ed25519 pack_signature; Loft Proof requires it — recomputing the pack's unkeyed hashes without the signer key does not verify. Proof is signer-anchored (fix round 2, D2X): the embedded key's self-consistency never suffices — Proof cross-checks signer fingerprint + embedded key against a verifier-supplied local registry / expected fingerprint (§7.1), and an attacker-minted self-consistent pack refuses UNVERIFIED_SIGNER.
Powered by Cyan Eclipse Technologies — https://cetech.ai