dev-5090 uncensored inference
activeUncensored/abliterated inference marketplace. Local RTX 5090 + ATOM node. Pay-per-call in USDC on Base.
Settled via Coinbase.
Not seen in any discovery source since 2026-11-14.
- Transactions · 30d
- 105
- Volume · 30d
- $5.51
- Unique buyers · 30d
- 13
- Uptime · 30d
- 86.3%
- Latency p50
- 224ms
- Reported calls · 30d
- 82
Endpoints (27 live)
POST/v1/defi/gaps— DeFi / WWW rail-gap analyzer: takes a set of live rails (settlement, price, listing, custody, latency) and returns a state-chain ledger of the gaps between them. Each node inherits prior state; a node cannot run if its required prior state is missing. The product is the gap ledger, not a market statement. (0.12 USDC on Base)POST/v1/flow/chain— Flow-state security job chain: recon → scan → exploit → audit → report. Each step is a state node that must inherit the prior node's state; a skipped step is refused, not silently run. Runs the FVCE red-team binaries. Returns the accumulated state, per-step cost ledger, and the chain that actually executed. (0.5 USDC on Base)POST/v1/j-space— J-Space: the gap between sense (27B, fast) and mind (125B, deep). Returns both answers + the difference + a proof-of-model stamp. The product is the DIFFERENCE, not the answer. (0.08 USDC on Base)POST/v1/coder/completions— Uncensored coder completions, served by Qwen3.8-27B abliterated (Heretic) on a local RTX 5090 via vLLM. (0.03 USDC on Base)POST/v1/sql-explain— Explain a SQL query in plain English. Paste query in message. (0.01 USDC on Base)POST/v1/coder/test-gen— Generate tests for a code snippet (matches file's existing test framework). (0.03 USDC on Base)POST/v1/curl-to-code— Convert a curl command to Python/JS/Go. Specify target language in the message. (0.01 USDC on Base)POST/v1/coder/refactor— Refactor code while preserving behavior. Returns diff + rationale. (0.03 USDC on Base)POST/v1/index— Embed and index documents into a private vector collection (768-dim, cosine). documents=[{id,text,metadata}], namespace picks the collection. Returns documents, chunks, tokens and latency. (0.05 USDC on Base)POST/v1/agent/compose— Pay-to-compose orchestrator: pulls live data (weather), routes it through two local-model stages, then an independent judge verifies the result against the source data before it ships (one bounded self-correction on rejection). Returns the composed answer, the judge's verdict, and a per-step cost ledger. (1 USDC on Base)POST/v1/vision— [preview] Image understanding. Currently TEXT-FALLBACK ONLY — no vision model is loaded. Priced accordingly. (0.02 USDC on Base)POST/v1/regex-explain— Explain a regex in plain English. Paste the regex in message content. (0.01 USDC on Base)POST/v1/reason— Multi-step reasoning and planning, served by Qwen3.8-27B abliterated (Heretic) on a local RTX 5090 via vLLM. (0.05 USDC on Base)POST/v1/chat/uncensored— Uncensored general chat (Qwen3.8-27B Heretic). (0.01 USDC on Base)POST/v1/translate— Translate text into a target language (ISO code: es, de, fr, ja, pt, pl...). Optional domain and tone hints. Returns the translation, the model, usage and latency. (0.02 USDC on Base)POST/v1/regex-build— Build a regex from a natural-language description. Returns just the regex + one test string. (0.02 USDC on Base)POST/v1/summarize— Summarize a document without it leaving this machine. mode=bullets|paragraph|key-points, max_words caps the length. Returns the summary, the model that wrote it, token usage and latency. (0.05 USDC on Base)POST/v1/coder/security-audit— Security audit of code snippet (SQLi, XSS, auth bypass, secret leaks, unsafe deserialization). (0.1 USDC on Base)POST/v1/attest/model— Proof-of-Model: detect silent model substitution at any OpenAI-compatible inference endpoint. mode=fingerprint records what a seller serves today; mode=verify re-probes later and proves whether the weights changed. Deterministic, evidence included, no model used to judge. (0.1 USDC on Base)POST/v1/agent-task— Complete a full agentic task (plan → generate → verify). Longer running. (0.1 USDC on Base)
+7 more endpoints.
First seen · last seen · last active