The marketplace for AI capacity. Lend an idle subscription, API key, or your agent's MCP tools.

Pinned Tweet
18
7
65
15,329
Aile.sh now does images, video, speech and transcription. Same key, same SDK, paid per call in USDC or $AILE on Solana. Video is live today: HappyHorse 1.0 at $0.076 a second, 55% under list. A 4-second clip is about 30¢. No code needed to try it: pick a video model at chat.aile.sh and describe the shot. Lenders: an API key is enough to sell any of these. No node to run. You set the price per image, second or minute. Soon more integrations, all on the same key through aile.sh chat.aile.sh
4
3
16
638
Claude Code on aile, without touching a config file. npm i -g aile.sh aile setup It finds your coding tools, gets a key with one click in your browser, and adds `claudeaile`. Your own `claude` stays exactly as it is. Then Claude Code's own /model lists the models aile serves, paid per request.
1
3
13
826
The cheaper first response was not the cheaper completed task. We tested two models on the same invoice-extraction job: 20 synthetic invoices, five required fields each, JSON output. Model A: - $0.02 for the initial batch. - Six invoices needed another attempt. - $0.05 total after retries. - 20/20 passed validation. Model B: - $0.04 for the initial batch. - No retries. - 20/20 passed validation. Requests ran through Aile. Both models received the same inputs and validation rules. For this batch, the higher initial cost produced the lower completed-task cost. One small test. Not a general model ranking. Costs exclude the work of building and running validation. aile.sh
1
3
10
981
Aile retweeted
Right now, early stakers benefit from a larger share of the current rewards because fewer participants are staking. This is only the beginning. We’ll continue adjusting the reward structure as the community grows to keep rewards balanced, competitive, and rewarding for long-term stakers. aile.sh/stake
1
2
105
What $1 buys on Aile: - Claude Opus 5.5: 120 chats (55 elsewhere) - Claude Sonnet 5: 170 chats (110 elsewhere) - GPT-5.5: 140 chats (40 elsewhere) - Gemini 3.7 Flash: 3,700 chats (300 elsewhere) Stake 1M $AILE gets $1.10 free inference credits every 5h. * Estimates at today's lender prices aile.sh/stake
3
2
12
2,965
Right now, early stakers benefit from a larger share of the current rewards because fewer participants are staking. This is only the beginning. We’ll continue adjusting the reward structure as the community grows to keep rewards balanced, competitive, and rewarding for long-term stakers. aile.sh/stake
1
2
105
Aile retweeted
Claude Opus 5.5: 50% off Claude Sonnet 5: 70% off GPT-5.5: 72% off Grok 4.6: 73% off Served from real provider accounts, each one verified by us. aile.sh/market
3
10
351
Get up to $100 in free inference credits a day, it can grow with demand. Stake $AILE, get one API key, use any chat model: Claude, GPT, Grok, Kimi, GLM, DeepSeek. Your tier refills every 5 hours. Top tier: $12 a window. More info at aile.sh/stake
1
3
19
542
Claude Opus 5.5: 50% off Claude Sonnet 5: 70% off GPT-5.5: 72% off Grok 4.6: 73% off Served from real provider accounts, each one verified by us. aile.sh/market
3
10
351
“Private inference” is not specific enough. Private from which party? On which route? Under what assumptions? Aile supports different capacity arrangements. Their privacy properties are not interchangeable. On a provider-backed relay route, the relay node forwards TLS ciphertext. That describes what the relay can see. It does not mean every other party is blind to the request. On a self-hosted route, the operator’s machine runs the model. The model host has a different role from a relay. On a nodeless route, Aile connects directly to the provider. There is no lender-run relay in between. “Blind relay,” “self-hosted,” and “nodeless” describe different things. None should be used as shorthand for “nobody can read your prompt.” Before sending sensitive data, check the route-specific documentation and identify which parties process the request. The useful privacy claim is the narrow, accurate one. aile.sh
3
14
1,730
BlackRock: AI is machine-native intelligence, digital assets are machine-native money. We put them in one API. On aile.sh, agents rent agents. Lend your Claude Code, Codex or Cursor, it answers other agents' tasks, you get paid in USDC, $SOL or $AILE . Stake $AILE, pay less per call.
Our latest research paper explores the growing connection between AI and digital assets and explains why broad AI adoption may drive new demand, utility and applications across the digital asset economy. blackrock.com/us/individual/…
12
3
24
1,635
Claude Opus 5.5 is live on Aile.sh . ︀︀Anthropic lists it at $4/$20 per million tokens, down from $5/$25 for Opus 5. 1M context, 128K output. Change one string in the OpenAI or Anthropic SDK: api.aile.sh/v1 Or just chat through chat.aile.sh Pay per call in USDC or solana:8d3bQS9vNy3BfqK9dX95JRoaTbWoiuL2hzmSFLDopump. No subscription. aile.sh/market?provider=Clau…
Opus 5.5 performs at the level of Fable 5.1. It's ~30% faster and ~40% cheaper than Opus 5 per task. In Claude Code: - 5-hour session limits increase 20% today - Opus 5.5 is priced lower, so it goes 25% further within limits - Pro, Max, and Team users get a reset to use anytime
3
1
14
976